The Political Economy of Local Democracy: Decentralization and Performance in Mozambique Since 1998

THE POLITICAL ECONOMY OF LOCAL DEMOCRACY: DECENTRALIZATION AND PERFORMANCE IN MOZAMBIQUE SINCE 1998

Beatrice Anne Reaud

Submitted to the

Faculty of the School of International Service

of American University

in Partial Fulfillment of

the Requirements for the Degree

of Doctor of Philosophy

International Relations

Chair:

Professor Deborah Bräutigam, American University

Assistant Professor Rachel Robinson, American University

Dr. Barak Hoffman, Georgetown University

Dean James Goldgeier

Beatrice Anne Reaud

2012

PENDING

To my dear son.

May your heart always be filled with wonder about the world around you and knowledge of how deeply you are loved.

THE POLITICAL ECONOMY OF LOCAL DEMOCRACY: DECENTRALIZATION AND PERFORMANCE IN MOZAMBIQUE SINCE 1998

Beatrice Anne Reaud

ABSTRACT

Governments in weak states routinely fail to deliver services they promised to provide their citizens. In theory, decentralization is supposed to promote accountability at the local level and hence, improve municipal performance by encouraging public participation, building government capacity, and increasing political competition. Decentralization should improve performance because it gives civil society, voters, and fee payers an opportunity to hold local officials accountable by incentivizing them with political and revenue pressures. However, even in cases where decentralization has occurred, local municipalities vary in the extent to which they are effective suppliers of services.

The literature on decentralization and performance suggests that participation, resources and voting are three accountability mechanisms that affect municipal performance (Crook and Manor 1998;

Agrawal and Ribot 1999; Cheema and Rondinelli 2007). Mozambique, a post-conflict, low-income new democracy that implemented a major decentralization effort in 1998, provided a natural laboratory in which to investigate how these factors cause variation in municipal performance. I found that public participation was low and donor-dependent, municipalities still relied on central government transfers and donor funding, and a single party dominated state resources, which limited political competition and accountability.

Whereas theory predicted that these conditions would not foster accountability and incubate municipal performance, I found that donors, unearned income, mayoral leadership and political competition improved municipal performance in three cases. First, donors provided resources that capacitated public participation and funded services. Second, decentralization provided authorities with

incentives for municipalities to double revenue collections in the face of high aid and central government transfers. Third, mayoral leadership was critical in using resources and authorities to improve performance. Lastly, decentralization provided the political opening in a de facto one-party state for an opposition mayor who performed well to win re-election and launch a new national party.

These findings extend beyond international relations to the fields of development studies, public administration and political science. Even in the context of weak institutions and serious constraints, local governments are able to build revenue capacity, improve service delivery, and sustainable political independence in dominant-party states through decentralization reforms.

ACKNOWLEDGEMENTS This research would not have been possible without the intellectual guidance and support of my dissertation committee, Rachel Robinson at American University, Barak Hoffman at Georgetown University and my special thanks to my chair, Deborah Bräutigam, who provided the leadership throughout the project to help me bring it to its concluding phase.

I would also like to especially acknowledge the mentorship of Peter Lewis at Johns Hopkins Paul H. Nitze School for Advanced International Studies who helped me procure initial funding, Bernhard Weimer and Jose Macuane at Eduardo Mondlane University in Maputo, Adriano Nuvunga at the Center for Public Integrity in Maputo and Ed Connerley. Their advice throughout the project helped shape my thinking and I am truly grateful for the experience, invaluable insights, and good humor.

Funding from the National Security and Exchange Program and American University made fieldwork in Mozambique from April 2008 to April 2009 possible as well as follow-up support for the writing of the dissertation. Field work in Mozambique would not have been possible without the help of Ali Alwahti, Uri Raich, Aparna Kumar, Signe Fredericksen, and Paulo de Renzio and the support of Eduardo Mondlane University in Maputo, CEDECA at the Pedogogic University in Beira, and Catholic University of Beira. I especially thank the openness of the municipal staff in Maputo, Dondo, Beira, and Matola for their time in answering my questions and Mayors Comiche of Maputo, Simango of Beira, and Cambezo of Dondo for their time and assistance. A special debt of gratitude goes to my tireless and resourceful research colleagues Calisto Bila, Ricardo Muchanga, and Ascenção Machel.

To my family, for instilling curiosity, desire, and aptitude to understand the world around me.

To my loving husband for support, encouragement, patience, humor and companionship in our life¶V journey. And, to our son, for filling my heart with joy and contentment.

(exchange rate used was from Sept. 2008 (26.5 MTn: USD$1))

TABLE OF CONTENTS

I. INTRODUCTION: IN THE DECENTRALIZED CONTEXT,

WHAT CAUSES VARIATION IN MUNICIPAL PERFORMANCE? 11

Introduction 11

Significance and Contributions 14 Institutional Inheritance 20 Explanatory Variables 32 Dependent Variable: Municipal Performance 39 Plan for Dissertation 43 Figure 1.1 Hypothesis 45 Table 1.1 Factors used to select municipal cases 46 Table 1.2: Municipal Performance and Explanatory Variables for Selected Cases 47

II. GROUNDING EMPIRICS IN THEORY 48 Introduction 48 The Hypothesis 49 Decentralization and Performance 50 Civil Society: External to and Extension of the State 51 Revenue: Earned and Unearned Income (Rents) 58 Elections: Democratic Performance versus the Performance of Democracy 62 Conclusion 67 Table 2.1. Hypothesis and Findings 69 Table 2.2. Theoretical Launching Point, by Municipality, Service and Variables 70

III. USING EXPERIMENTAL APPROACHES: DEMONSTRATING VARIATION ON THE DEPENDENT VARIABLE 71 Introduction 71 Initial Evidence: Summary of Independent Variables by Case 72 Ambulatory Vendor Experiment 77 Summary of Ambulatory Vendor License Experiment 81 Trash Observation Experiment 83 Conclusion 94 Table 3.1. Expected versus Actual Performance 95

Table 3.2. Ambulatory License Process 96

IV: ESTAB/,6+,1*&,9,&3$57,&,3$7,21¶6EFFECTS ON MUNICIPAL PERFORMANCE 97

Mozambican Civil Society: Current State and Departure Point 100

Analysis by Case 106

Maputo 106

Beira 117

Dondo 129

Matola 135

Conclusion 139

Table 4.1. Hypothesis and Finding 143

Table 4.2. Expected versus Actual Findings 144

V: REVENUE EFFECTS ON SERVICE DELIVERY 145

Introduction 145

Overall Trends: Revenue Flows 149

Across time, within case 157

Table 5.1. Maputo 157

Table 5.2. Beira 158

Table 5.3. Dondo 160

Table 5.4. Matola 162

Performance Analysis 163

Conclusion 174

Table 5.5. Expected v. Actual Outcomes 177

Table 5.6. Across Case Averages: 2004-2007 178

Table 5.7. Selected aid projects, 2004 to 2007 179

Table 5.8. Conclusion 181

VI: VOTING EFFECTS ON SERVICE DELIVERY 182

Introduction 182

2003 Mayoral Elections 185

Performance from 2003 to 2008 190

2008 Mayoral Elections 199

Conclusion 206

Table 6.1. Hypothesis and Findings 211

Table 6.2. Expected versus Actual Findings 212

Table 6.3. Electoral Turnout 213

Table 6.4. 2003 Municipal Elections 214

Table 6.5. 2008 Municipal Elections 215

VII. CONCLUSION: INSIGHTS AND PATHS OF INQUIRY 216

Introduction 216

The Hypothesis, Revisited 218

Civil Society and Municipal Performance: Strategic Influence 219

Revenue and Municipal Performance 224

Elections and Municipal Performance: Contracting and Expanding Spaces 228

Leadership and Citizenship on Municipal Performance: Factors to Consider 233

Conclusion 237

Table 7.1. Hypothesis and Findings 240

Table 7.2. Expected v. Actual Findings 241

Bibliography 242

I. INTRODUCTION: IN THE DECENTRALIZED CONTEXT,

WHAT CAUSES VARIATION IN MUNICIPAL PERFORMANCE?

'HFHQWUDOL]DWLRQUHIRUPV«ZLWKWKHSURFHVVRISROLWLFDOWUDQVLWLRQLQWKHV«EHFDPH DVVRFLDWHGZLWK«DQHPHUJHQFHRIQHZDFWRUVPRELOL]DWLRQRIµFLYLOVRFLHW\¶FRQVWUXFWLRQRID local political playing field and a renewal of participatory practices. Nevertheless « LPSOHPHQWDWLRQ«has still not had a significant impact LQWKH«capacity to respond through local government to the necessities of residents and improve the conditions of their lives (Forquilha 2007, 1-2).1

Salvador Cadete Forquilha, Institute of Social and Economic Studies, Maputo, Mozambique (emphasis added, September 19 , 2007).

In 1998, we died like ants. Due to the Mayor, now we see streets paved and things are starting to look normal. Although we may not have trash collection where I am, it is clear that this mayor is working and that things have gotten better. Focus group participant, Beira, Mozambique, March 14, 2009.

Introduction

In the euphoria created by the democratization wave in the 1990s in developing and transition

FRXQWULHVVFKRODUVKHUDOGHGWKH³HQGRIKLVWRU\´DVGHPRFUDWLFJRYHUQPHQWVSUHDGWRLPSURYHWKH quality of life for residents all over the world (Huntington 1991; Fukuyama 1993). Decentralization has been embraced by donors and recipient governments alike as a way to, in theory, strengthen nascent democratic governments in weak states by promoting accountability and better service delivery because local governments are closer to the citizens they serve and citizens are empowered to hold public officials accountable (Connerley 2009; Cheema and Rondinelli 2007; Bräutigam 2004; Bird 2000; Bird and Vaillancourt 1998). As a result, implementation of decentralization programs was, and continues to

1 This sentiment is echoed by Dr. John Barnes with National Directorate of Planning and the Budget ZKRLQFOXGHGLQDSUHVHQWDWLRQWKDW³UHDOGHFHQWUDOL]DWLRQLVYHU\OLPLWHGLQ0R]DPELTXH6HH'U-RKQ %DUQHV³Descentralização em Moçambique: Mito our Realidade?´3RZHUSRLQWSUHVHQWDWLRQ&7$ Program of Planning and Decentralized Finance, Directorate of National Planning and the Budget, Maputo, Mozambique.

be, an integral component of democratization strategies in the developing world, despite mixed results as the above-stated quotes suggest.2

The literature on decentralization focuses on macro-processes, likening implementation of successful refoUPVWRSUHSDULQJD³VRXIIOp´PHDQLQJLWUHTXLUHVDSUHFLVHDQGFRXQWU\-specific sequencing of devolution of fiscal, political and administrative authorities to alter institutions in favor of increased economic development, political representation and service delivery.3 However, seemingly little attention is paid to addressing a crucial question: what factors at the micro-level influence municipal performance?4 Specifically, within cases in which decentralization has occurred and institutional factors such as historical context, elections and revenue authorities remain constant, local municipalities vary in the extent to which they are effective suppliers of local services. What accounts for this variance?

The purpose of this dissertation is to answer that question. My hypothesis was that decentralization will promote accountability at the local level and hence, improve municipal performance, by encouraging public participation, resource generation, and political competition. In other words, decentralization should improve performance because it gives civil society, voters, and fee payers an opportunity to hold local officials responsible to providing high municipal service by incentivizing them with political (votes, lobbies, and increased citizen participation) and revenue (fees and taxes) pressure (see Figure 1). More succinctly, decentralization creates incentives that motivate government officials to act and opportunities for residents to pressure them.

However, this framework assumed that accountability is not distorted by dependence on aid, low per capita income, and lack of political competition that are realities in many weak states. So, what

2 Mozambican scholar Fedilx Kulipossa concluded WKDW³GHFHQWUDOL]DWLRQLVDFRPSOH[DQGPXOWL- IDFHWHGSKHQRPHQRQWKDWPD\KDYHERWKSRVLWLYHDQGQHJDWLYHHIIHFWV´ .XOLSossa 2004, 768)(Cheema DQG5RQGLQHOOL %RWK&URRNDQG0DQRU¶VDQG$JUDZDODQG5LERW¶VZRUNVUHVSHFWLYHO\ reference World Bank and USAID reports on promoting decentralization (See pp. 2-3 in Crook and Manor and footnotes 2 and 7 in Agrawal and Ribot) (Crook and Manor 1998)(Agrawal and Ribot 1999). 3 $QGUHZ3DUNHUDVTXRWHGLQ*6KDEELU&KHHPDDQG'HQQLV$5RQGLQHOOL³)URP*RYernment 'HFHQWUDOL]DWLRQWR'HFHQWUDOL]HG*RYHUQDQFH´Decentralizing Governance: Emerging Concepts and Practices, Washington, D.C. Brookings Institution, 2007: 9. 4 :HNZHWHPDNHVWKHOLQNH[SOLFLW³'HFHQWUDOL]DWLRQUHTXLUHVORFDORUJDQL]DWLRQ´ &KHHPDDQG Rondinelli 2007, 264).

accounts for variation in municipal performance in a country like Mozambique, which is subject to these types of resource constraints and weak institutions? I hypothesized that, in this context, weak citizen participation that cannot pressure the government to provide better services, municipalities have little resources and are highly dependent on aid and central government transfers, and figurehead mayors are subject to very little voter scrutiny because Mozambique is a de facto one-party state. My research found that these conditions are indeed present. Across all four cases, civil society was generally weak, municipalities relied on central government transfers and donor project funding, and a single party dominated state resources.

I found some interesting variation in the presence of these conditions that requires some rethinking of the current theoretical debates. First, civil society organizations are fluid and adaptive to resource availability. While not a pre-requisite to improving municipal service provision, in two cases, civil society provided a venue for participation and voice that led to improvements in municipal service in the form of protests and lobbying for city contracts in Maputo and participatory budgeting in Dondo.

Donor aid helped build technical capacity in both cases. I also found in Beria, Dondo and Maputo vendor association that acted are interlocutors between the municipality and vendors at a minimum and, in two cases, performance municipal functions. Second, own-source revenue grew in the presence of majority shares of unearned income. Third, regardless of party, mayoral leadership improved municipal service delivery. Lastly, decentralization provided the political opening in a de facto one-party state for the opposition mayor to both perform well and have a greater chance of being rewarded with re-election than dominant-party mayors. I discuss each of these findings in the chapters to follow.

This chapter is divided into five sections. First, I address why studying decentralization and performance is significant. Second, I outline why Mozambique is the ideal laboratory in which to explore variation in municipal performance. Third, I explore how the inheritance of institutions² namely colonialism and decentralization²set the stage for municipal performance in Mozambique. I then discuss my methodology²embedded case design²and how I chose the four Mozambican

municipal cases and services that I used to test the hypothesis stated above. Lastly, I conclude the chapter with a review of the dissertation, proceeding to the theoretical framework in Chapter Two.

Significance and Contributions:

Why study performance in the context of decentralization?

Significance

From an empirical perspective, the study of municipal performance in the context of decentralization is significant for three reasons: first, the number of people living in urban spaces globally has grown; second, the increasingly prevalent mode of governing people is through subnational government; and third, urban spaces tend to be economic drivers thereby fostering economic development. 7RGD\PRUHWKDQKDOIRIWKHZRUOG¶VSRSXODWLRQUHVLGHVLQFLWLHV In 1950, the urban population of the world was 30 percent (Cheema 2007, 171). By 2000, it expanded to 47 percent and is expected to grow to 70 percent by 2050 (Ibid). 5 The number of subnational governments has also multiplied. In 1980, only 10 of the 48 largest countries in the world had elected subnational governments (Eckardt 2008, 2). By 2000, the number grew to 34. Two-thirds of sub-Saharan African countries alone have recently implemented decentralization programs (Aworti 2010, 620). Lastly, cities and towns drive economic growth. One World Bank study reported that, although income growth in

$IULFDPD\EHVORZLQFRPSDULVRQWRRWKHUUHJLRQVWKH³LQFRPHRIVOXPGZHOOHUVHYHQWXDOO\LQFUHDVHV

ZLWKJURZWKDQGWKHVOXPVDUHXSJUDGHGDQGWXUQHGLQWRPLGGOHFODVVQHLJKERUKRRGV´ )D\DQG2SDO

2000, ,Q0R]DPELTXHDORQHWKHFHQWUDOJRYHUQPHQW¶VFKLHIVRXUFHRIUHYHQXH DVLGHIURPIRUHLJQ

5 3OHDVHDOVRVHH866WDWH'HSDUWPHQW³3UHVV5HOHDVH86'HSDUWPHQWRI6WDWH,QYLWHV863ULYDWH 6HFWRU(QJDJHPHQWLQ:RUOG8UEDQ)RUXP´%XUHDXRI:HVWHUQ+HPLVSKHUH$IIDLUV-DQXDU\ 2010, accessed 1/27/2010.

aid) is generated by the economic activity in municipalities. One long-time follower of Mozambican

GHFHQWUDOL]DWLRQLQGLFDWHGWKDW³SHUFHQWRIUHFHLSWVIor the central government come from the

HFRQRPLFDFWLYLW\JHQHUDWHGE\PXQLFLSDOLWLHV´6

From a theoretical perspective, the study of decentralization and performance is significant in the context of the study of institutions because it addresses how institutions change outcomes. Mainstream

LQWHUQDWLRQDOUHODWLRQVWKHRULVWVGHILQHLQVWLWXWLRQVDV³SHUVLVWHQWDQGFRQQHFWHGVHWVRIUXOHV´WKDW prescribe behavior (Keohane 1998, 383).7 Decentralization is fundamentally a set of policies, which modify rules of interaction between national and subnational governments by deconcentrating, devolving or decentralizing political, administrative or fiscal authorities. The theory goes that decentralization reforms are changes to institutions with the end goal of making government more inclusive, democratic, and accountable. So, for example, elections as an institution may be introduced under a decentralization rubric so that local officials who were originally appointed by the central government are elected with the goal of improving local accountability in service delivery. The result is that the behavior of central government officials changes because, in theory, they are now counterbalanced by locally elected officials.

Lastly, tying both practical and theoretical considerations together, decentralization is thought to improve performance because it increases accountability by changing institutional incentives and opportunities. Specifically, theory posits that decentralization of taxation and election authorities increases accountability because it creates both incentives for local officials to keep their newfound control over local resources and opportunities for new groups to gain access to and share control of those same resources. Local officials are incentivized by re-election so voters can hold them accountable by withdrawing or providing votes. Similarly, fee payers can hold local officials

6 Excerpt from the presentation of Dr. Abdul Illal on April 16, 2008 in Maputo, Mozambique at the National Seminar on Decentralization, April 16-April 18, 2008 sponsored by the Ministry of State Administration and the World Bank. 7From a public policy pHUVSHFWLYHGHFHQWUDOL]DWLRQLVGHILQHGDVWKH³WUDQVIHURIDXWKRULW\UHVRXUFHV DQGSHUVRQQHOIURPWKHQDWLRQDOWRWKHVXEQDWLRQDOOHYHOV´ 1GHJZD

accountable by withholding fees needed to fund the services on which local officials rely to justify re- election. This change in the incentives of local officials and the opportunities to pressure them should lead to responsive, democratic government²a key theoretical concern of political science.

Contributions

As discussed above and will be further expanded upon in Chapter Two, decentralization is a popular public policy reform adopted by donors and countries alike to improve government performance because it is expected to lead to greater accountability. Decentralization observers Cheema and Rondinelli stated:

«GHFHQWUDOL]DWLRQUHPDLQVDFRUHSUHVFULSWLRQRILQWHUQDWLRQDOGHYHORSPHQWRUJDQL]DWLRQVIRU promoting democratic governance and economic adjustment and is seen by many of its advocates as a condition for achieving sustainable economic, political, and social development DQGIRUDWWDLQLQJWKH81¶V0LOOHQQLXP'HYHORSPHQW*RDOV &KHHPDDQG5RQGLQHOOL, 10).

Despite its popularity, there is a limited body of empirical research that analyzes how decentralization actually works in practice and the specific factors that affect municipal performance to indicate that decentralization as a development and democratizing tool merits the investment it currently receives.

The literature on decentralization is abundant with cross-national studies and abstract models that suggest varying degrees of success and failure. Some scholars suggest that decentralization can result in the co-opting of local resources, bureaucrat bribes, under-provision of services and predatory central government (Bardhan 2002; Bardhan and Mookherjee 2006; Weingast 2009). Still, other studies suggest that, under certain circumstances, decentralization is a means to economic development (Bahl and Linn

1994). Mozambican decentralization scholar Fidelx Kulipossa concludes that additional evidence of the

HIILFDF\RIGHFHQWUDOL]DWLRQ¶VSHUIRUPDQFHLVZDUUDQWHG+HVWDWHV³«IXUWKHUWKHRUHWLFDODQGHPSLULFDO research is needed to ascertain whether there is a causal relationship between decentralization and

GHPRFUDF\´ .XOLSRVVD, 778). Additionally, accountability as a concept presumes that voters/tax payers, civil society groups and local officials are incentivized in the way theory suggests. However, in

an environment of scarcity²scarcity of political competition and resources²the ability for voters and fee payers to hold local officials accountable could be weak. Moreover, even if they could, the capacity for local officials to institute change²assumed in an accountability framework²remains in question.

My dissertation explores theories related to how different pressures²namely revenue and political pressures²hold municipal governments accountable to improve their services by providing evidence of how municipal performance improved in a case where the conditions²weak participation, low resource capacity, and weak political competition²would lead one to think that improvement in municipal services was not likely. Mozambique presents an excellent opportunity to examine the factors that explain variability in performance, given these constraints and the institution of decentralization reforms, which permit collection of in-country empirical evidence of municipal variation in performance. In subsequent chapters, I offer evidence that unearned income actually catalyzed improvement in services in two ways. First, aid built the capacity of public participation in a sustainable way that led to improvement in municipal performance in two cases. Second, contrary to the prevalent view that unearned income reduces the incentive for government to pursue own-source fee collection, decentralization provided authorities and incentives for municipalities which they used to double their own-source revenue collections despite high aid and central government transfers. Third, I found that, while participation was able to improve municipal performance in cases where donor aid helped develop technical capacities and municipalities worked with these organizations and processes, strong civil society was not a pre-requisite to improved performance. I also found that citizen protests in one case effectively challenged and modified municipal policy in the collection of trash fees. My research also addressed how, regardless of political party, mayoral leadership was critical in using resources and authorities to improve performance. Lastly, as previously discussed, decentralization provided the political opening for an opposition mayor who performed well to be rewarded not only with re-election but also political support to start a new, national political party. Mozambique is not unique in the constraints it faces. As such, these findings extend beyond Mozambique to countries whose residents similarly grapple with poverty and weak accountability mechanisms.

Why study performance in the context of decentralization in Mozambique?

As I stated in the previous section, decentralization has also been framed as a means to promote economic development and democratic change, provide ethnic minorities with representation and diffuse ethnic conflict.8 An ever-JURZLQJVKDUHRIWKHZRUOG¶VSRSXODWLRQLVOLYLQJXQGHUGHFHQWUDOL]HG government. The African continent, home to almost 1 billion individuals currently and projected to reach 2 billion by 2050, or roughly 20 percent of the global population, is the fastest growing continent

(Fengler 2010). Two-thirds of Sub-Saharan African countries (30 out of 47 countries) have enacted decentralization measures (Ndegwa 2002).9

At first glance, Mozambique has much promise as a case study in decentralization. As of 2007,

Mozambique had one of the highest proportions of urban populations in East Africa, with 36 percent of

Mozambicans living in urban areas. It has held peaceful national and local elections since 1994 after a

16-year civil war. Mozambican elected leaders have been recognized as effective by members of the international community.10 Despite the global recession, the Mozambican economy has consistently grown over the past 15 years, (The Economist, $IDOWHULQJSKRHQL[ ,WKDV$IULFD¶VPRVW

SRZHUIXOK\GURHOHFWULFSODQWLQ&DKRUD%DVVDGDPDQGSRWHQWLDOO\WKHZRUOG¶VODUJHVWXQH[SORLWHG coalfield in Tete province.

8 An excellent 30-country cross-national study examines the idea that decentralization alleviates ethnic conflict and found that, in cases where regional parties become dominant, it actually increased ethnic FRQIOLFW6HH'DZQ%UDQFDWL³'HFHQWUDOL]DWLRQ)XHOLQJWKH)LUHRU'DPSHQLQJWKH)ODPHVRI(WKQLF &RQIOLFWDQG6HFHVVLRQLVP"´International Organization, Vol. 60, No. 3 (Summer, 2006): 651-685. 9 This survey excluded island nations, which, if included, would have brought the number of countries from 47 to 53. 10 7KHILUVWSUHVLGHQWHOHFWHGWKURXJK0R]DPELTXH¶VILUVWPXOWLSDUW\HOHFWLRQVLQ, President Joaquim Chissano was awarded $5 million after he stepped down from his second term as President by the Mo Ibrahim Foundation, a new foundation dedicated to improving African governance. From a pool of 13 leaders Mr. Chissano was chosen based on his ability to broker an end the crippling 16-year civil war in Mozambique, shepherd a new Constitution based on democratic principles and transitioning a primarily socialist economic to a market-based one.

However, despite these indicators, Mozambique is one of the poorest and most aid-dependent countries in the world, remaining an effective one-party state, with international observers pointing to the ever-growing fragility of this new democracy. In 2010, international aid is expected to total

USD$1.60 billion, half of the total state budget.11 $FFRUGLQJWRWKH,0)0R]DPELTXH¶V*'3SHU capita is USD$1,000 per year, landing it in the bottom twelve poorest countries in the world (Ibid).

Despite a 15 percent decline in the poverty rate from 69 percent in 1997 to 54 percent in 2003, the absolute number of those living below the poverty line rose (Ibid). According to the UNDP 2007/2008

Human Development Report, Mozambique ranked 172 out of 177 countries with data on factors such as life expectancy and adult literacy.12 Price increases to basic commodities and services, like bus fares and bread in February 2008 and September 2010 respectively in a country with virtually no safety net, were met with angry rioters and heavy-handed government responses.

In terms of governance reforms and democratic consolidation, Mozambique also seems to be wavering. An international observer of democratic performance, Freedom House, dropped

Mozambique from its list of electoral democracies as a result of electoral issues surrounding the 2009 national and provincial elections (Ibid). Mozambican politicians point to a crisis in leadership and corruption in a de facto one-party state. ³2XUSUREOHPLQ0R]DPELTXHLVOHDGHUVKLS´VWDWHV$IRQVR

'KODNDPDRSSRVLWLRQOHDGHURI5(1$02³0R]DPELTXHKDVYDULRXVQDWXUDOUHVRXUFHVULFKVRLOIRU agriculture, much water, mineral resources. However, 65 percent of the national budget depends on extHUQDODLG´ -&KDPXVVH $ORFDOMRXUQDOLVWDGPLWWHGWKDWWKH³WHQWDFOHVRIWKHUXOLQJSDUW\DUH

LQDOPRVWHYHU\VHFWRURIWKHHFRQRP\´ 1RWTXLWHDV6WHOODUDV,W/RRNV ,QFRQFOXVLRQ

11 However, it is not simply the quantity of aid but how it displaces capacity. Donors spend $350 million a year hiring 3,500 technical consultants, which could pay 400,000 local civil servants. See The (FRQRPLVW³$IDOWHULQJSKRHQL[0R]DPELTXH¶VUHFRYHU\´0LGGOH(DVWDQG$IULFDVHFWLRQ86 edition, July 10, 2010, accessed from Lexis Nexis Academic on December 28, 2010; The Economist, ³$VFUDPEOHLQ$IULFDWKHIXWXUHRIDLG´,QWHUQDWLRQDOVHFWLRQ6HSWHPEHU$FFHVVHGRQ December 27, 2010 through LexisNexis Academic. 12 The HDI examines three aspects of human development: life expectancy, adult literacy and enrollment in school, and purchasing power parity. According to the 2007/2008 UN Human Development Report, Mozambique ranked 171 on life expectancy (42.8 years), 131 on literacy (literacy rate was 38.7 percent), and 154 on GDP per capita (1,242 USD). Human Development Report 2007/2008 Country Fact Sheets-Mozambique 2008.

0R]DPELTXH¶VFRQVWUDLQWVPDNHWKHVXFFHVVHVHYHQPRUHXQexpected, thereby providing evidence that government performance can improve in the presence of institutional reform even in the most tenuous of circumstances. Critical to the study of variation in performance is the understanding of what remains constant. In the next section, I explore four institutional conditions, which remain constant throughout

Mozambique and affect all municipalities in order to be able to study variation in municipal performance.

Institutional Inheritance: Colonialism

And Socialism, Destabilization And Decentralization

Colonialism and Socialism

Located at the nexus of Arab, Indian, and Portuguese trade routes, Mozambique, roughly the size of France and Spain combined, today is home to 20 million people from European, South Asian, and various indigenous groups. These groups include the Shangaan in the South, Makonde and Makua in the North and Sena and Ndau in the Center who speak almost 20 native languages in addition to the national language of Portuguese (Mazula 2004, 183-200; Census 2007). Prior to the arrival of the

Portuguese in the 15th century, Bantu-speaking people migrated to present-day Mozambique, married into the local indigenous groups and created communities organized into independent chieftaincies governed by land chiefs. Gold and ivory attracted Arab traders who established a strong presence in

Northern Mozambique and along the coast. The Portuguese eventually challenged them in 1525resulting in their establishment of control in Mozambique from a series of coastal capitals by providing land grants to European settlers. Portuguese hegemony reached its apex by the 17th century and gradually

ceded control to landholders, Muslim traders, Indian merchants and later, British and other European concessionaries. By the 18th century, trade shifted from commodities like gold and ivory to slave labor and later was followed by export commodities such as timber and shrimp.

The Portuguese brought with them and implanted a system of government and of territorial administration, which was taken over by FRELIMO (Frente de Libertação de Moçambique) at

Independence and still in use today. The need for the colonial state to exercise central control over its

Mozambican territory and its people resulted in administrative structures such as District Administrators

(for rural areas) and boards, such as Câmaras Municipais, which were distinctly Portuguese. According

WR0R]DPELFDQODZSURIHVVRU*LOOHV&LVWDF³7KHFRXQWU\LQKHULWHGIURPWKHFRORQLDOSDVWDQ administrative structure essentially based on the principal of centralization, that is, the centralization of

DGPLQLVWUDWLYHGHFLVLRQVLQWKHVXSHULRURUJDQVRIWKHFHQWUDODGPLQLVWUDWLRQ´ &LVWDF, 32).

Mozambican scholar Joseph Hanlon observed how FRELIMO adopted this system:

FRELIMO adopted a government structure little different from that of the Portuguese. The president was at the top and he appointed governors and ministers. Below that were appointed national and provincial directors and district administrators. The structure of the national government was replicated at provincial and district levels, with directors of health, education,etc. But governors, district administrators and provincial and district directors had little previous experience and rarely received any special training (Hanlon 1991, 14).

A highly-centralized, bifurcated governance system arose in which districts were divided between

European (urban) and non-European (rural) areas. European areas were the townships or conselhos,

ZKLFKOLNHWRGD\¶VPXQLFLSDOLWLHVHQMR\HGVRPHGHJUHHRIVHOI-government in towns like Laurenço

Marques (Maputo) and Beira through the Municipal Board (Câmara Municipal)7RGD\¶VGHFHQWUDOL]HG municipalities are known as municipal councils (Conselhos Municipais). The indigenous population lived outside this system in rural areas, which were subdivided into posts (postos) headed by a chief

(chefe do posto), a structure that still exists today.13

13 The Portuguese appointed and paid régulos or members of the indigenous ruling classes to collect taxes and arbitrate local disputes. Essentially two segregated forms of order evolved: one formal in

After a ten-year war for independence, Portuguese colonial control ended in 1975 followed by a period of single-party dominance by FRELIMO. When FRELIMO assumed power, major ethnic groups from the North and Center were marginalized as a result of internal politics. Ironically, FRELIMO made use of the strong centralized government inherited from the Portuguese mixed with a Marxist-

Leninist ideology that sought to modernize the country. Traditional and tribal courts were abolished and régulos appointed by the Portuguese lost their official status. FRELIMO adopted a philosophy of

³GHPRFUDWLFFHQWUDOLVP´LQZKLFKWKHSDUW\DQGVWDWHPHUJHGDQGWKHSDUW\UHPDLQHGFRQQHFWHGWRLWV social bases to respond to popular demands. The party/state system still extended down to the neighborhood, or bairro, level, where there are block chiefs, chefes do quarterão, and secretaries of the bairro.14 This structure was imposed by the Portuguese government and, in practice, the FRELIMO party did little to change it.

Five years after the Portuguese left Mozambique, the country entered another phase of struggle.

Between 1977 and 1992, Mozambicans lived through a harrowing guerrilla war waged between the

Marxist-Leninist leaning FRELIMO government and the MNR (Movimento Nacional de Resistência), which eventually became RENAMO (Resistência Nacional Moçambicana). It is estimated that destruction totaled USD$15 billion and that 900,000 Mozambicans were affected, with Tete and

Zambezia provinces being the most affected (Abrahamsson and Nilsson 1995, 66). RENAMO

ZKDWEHFDPHWKHXUEDQDUHDVIRU(XURSHDQVDQGRQHEDVHGRQWKHFRORQLDODGPLQLVWUDWRUV¶DQGUpJXORV¶ LQWHUSUHWDWLRQVRI³WUDGLWLRQDO´ODZRXWVLGHXUEDQDUHDVIRUWKH indigenous population. Black Mozambicans overwhelming lived in rural areas, which were subdivided into postos. The chefe do posto ran these areas. This structure still exists today and is pronounced in cities like Matola. The legal system, inherited by the Portuguese, was highly formalized, with bureaucratized procedures, which remain in place today as is evidenced in many of the steps citizens take to get something as simple as an DPEXODWRU\YHQGRUV¶OLFHQVH7KHUROHRIvereador, which is like a town councilor, is a Portuguese institution. A bifurcated legal system segregated Europeans from the indigenous residents in which the former lived under colonial law, rooted in Portuguese metropolitan law, and the latter was subject to ³WUDGLWLRQDO´OHJDO systems as they were understood and applied by colonial administrators. Characteristic of colonial law was an emphasis on highly formalized procedures and paperwork, a practice, which persists today. 14 To do interviews today of residents, researchers must receive permission still from the secretary of the bairro and in at least one case city, the secretaries of the bairro are often politically connected and helped the incumbent mayor get re-elected in another city, a secretary of the bairro was nervous about doing focus groups during an election for fear of swinging the vote away from the incumbent.

eventually became a legitimate political party after the Peace Accords, signed with the FRELIMO government in Rome on October 4, 1992.15 The first preVLGHQWHOHFWHGWKURXJK0R]DPELTXH¶VILUVW multiparty elections in October 1994, President Joaquim Chissano, won by 53.3 percent of the vote, with 129 seats in the National Assembly won by FRELIMO. The RENAMO candidate, Afonso

Dhlakama, garnered 33.7 percent of the vote, with his party winning 112 seats (Vines 1991). The

Constitution was based on democratic principles and transitioning a primarily socialist economy to a market-based one. However, despite these indicators and a history of reconciliation, Mozambique still remains a de facto one-party state.

The war of destabilization had two important effects relevant to this study: it both increased the centralizing tendencies of the state and created the two-party landscape that exists today which affects municipal performance.16 All 11 provinces experienced the war to some extent. Movement among urban areas was tightly controlled by the central government as a result. The war also had a major effect on the growth of municipalities because rural dwellers migrated to urban centers to escape

RENAMO raids.17 The war fueled urban informal growth that still causes major impediments to urban development today in providing services to sprawling informal communities. In all case studies, informal settlementVVSURXWHGLQZKDWLVWRGD\WKHFHQWHURIWKHRULJLQDORU³FHPHQW´FLW\NQRZQDV

15 The main opposition party is actually a coalition between RENAMO and a series of smaller political parties, which together are known are RENAMO-UE (União Eleitoral). However, the short-hand RENAMO is used throughout the report. 16 The war of destabilization has had several effects of Mozambique which have been well documented by scholars such as Joseph Hanlon, Allen and Barbara Isaacman, Alex Vines, Harry West, João Cabrita, among others and annual reports by the UNDP, among other international aid agencies, which document the education and human development in Mozambique. These broader effects remain outside the focus of this project. For more information, please consult the annual Human Development Report and HDI index of Mozambique, which can be found on the UNDP website: http://www.undp.org/. For more details on the effects of socialism and the war of destabilization on Mozambique, please see: Joseph Hanlon, Mozambique, Revolution Under Fire, London: Zed Books, Ltd, 1990, and Mozambique, Who Calls the Shots? London: James Currey, Ltd, 1991, Harry G. West, Kupilikula: Governance and the Invisible Realm in Mozambique, Chicago: University of Chicago Press, 2005, Allen and Barbara Isaacman, Mozambique: From Colonialism to Revolution, 1900-1982, Boulder, Co,: Westview Press, 1983, Alex Vines, RENAMO: From Terrorism to Democracy in Mozambique? London: James Currey, Ltd, 1996, and João Cabrita, The Tortuous Road to Democracy1HZ

such because the buildings are formal and made from cement as opposed to the cinder blocks, reed or wood construction in informal communities. These settlements took the shape of nuclear villages, in which the elder male settled on a plot of land and then relatives came and lived surrounding this elder.

This form of cellular urban growth was highly disorganized and resulted in informal homes being built close together, making interior access for the provision of municipal services today difficult.

Decentralization

Decentralization was driven by two factors: post-war economic and political stabilization, which required the creation of political space for RENAMO, and the subsequent need for economic recovery and shift from a central, planned economy to a market-based economy. Motivated by economic and political instability and the advice of donors such as the World Bank, economic liberalization began in

WKHHDUO\VZLWKDFXOPLQDWLRQLQWKH0R]DPELFDQJRYHUQPHQW¶VDGRSWLRQRIDQ(FRQRPLF

Rehabilitation Plan in 1987. 18 Simultaneous to the adoption of market-opening reforms, donors used the opportunity to promote political liberalization. As the party in charge, FRELIMO proposed and passed a Constitution which, while locking into place conditions that were favorable to their continued dominance, both opened up political competition to the new opposition²RENAMO²and created a space for democratization of local government (Manning 2005). While local government elections created an opportunity for competition among political parties, it also placed the managerial capacity and leadership of now elected-mayors under scrutiny, a condition not present prior to decentralization.

Titles III and IV of the 1990 Constitution lay the foundation for decentralizatLRQWKURXJK³ORFDO

DXWKRULW\´ poder local) and defined the autarquias, which comprise the municipalities. This

18 Economic Rehabilitation Plan (PRE or the Plano de Reabilitação Económico) became the Economic and Social Rehabilitation Plan (PRES or Plano de Reabilitação Económico e Social ). A key component of this program was the gradual decentralization of authority to local governments. World Bank, Mozambique-Municipal Development Project, Project no. MZPE1806, Report no. PID8361, Appraisal date February 19, 2001.

FRQVWLWXWLRQDODPHQGPHQWZDVIROORZHGE\WKHHQDFWPHQWRIWKH³GHFHQWUDOL]DWLRQSDFNDJH´LQ

The same year, the first Presidential and Parliamentary elections were held and the Parliament passed

Law no. 3/94, which established elections for both rural and urban municipalities and their administrators and mayors, respectively. After FRELIMO had won the Parliament with only a 17-seat margin, tKHSDUW\¶VOHDGHUVKLSKDGVHFRQGWKRXJKWVDERXWH[WHQGLQJGHPRFUDWLFGHFHQWUDOL]DWLRQWRWKH rural areas, where RENAMO had its strongholds. This political calculation further solidified the bifurcated nature of Mozambican government: urban areas, which became decentralized and rural areas that experienced limited deconcentration.

Law no. 10/97 designated 33 municipalities, with an additional 10 added by Law no. 3/2008. The package of laws enshrines administrative, financial, and patrimonial autonomy for municipalities and defines their competencies. Although provinces and districts have experienced some delegation and deconcentration of central authority, the cornerstone of Mozambican decentralization remains the municipality. Although the first multiparty municipal elections were scheduled for 1996, they did not take place until 1998, after being postponed three times due to RENAMO-FRELIMO party politics

(Weimer and Fandrych 1999). RENAMO boycotted these elections, feeling that they were completely excluded from negotiations on the selection of municipalities that would be decentralized, with separate political and revenue authorities. This boycott resulted in only a 15 percent voter turn-out (Cuereneia

2001).

Today, Mozambique has four main levels of government:

1. Central government (consisting of the President, Prime Minister, Council of Ministers, National Assembly, and line and sectoral ministries); 2. 11 provincial governments (consisting of delegations of central government ministries and directorates); 3. 128 district governments (consisting of district-level central government delegations which provide services to non-decentralized municipalities and villages and administrative posts, which act as bureaucratic subdivisions); and 4. 43 autarquias (or municipalities, which are politically autonomous units).

The central government is a multiparty, semi-presidential system in which the President of the

Republic is elected through direct universal suffrage to a five-year mandate and appoints the Prime

Minister who is accountable to the President rather than the 250-seat National Assembly. The President presides over the Council of Ministers, convened by the Prime Minister. The National Assembly is elected according to a system of proportional representation under closed party lists. The FY 2010 State

Budget totals MTn118 mil milhões (USD$3.90 billion), which funds the activities of all ministries and various levels of the government.

There are 11 provincial and 128 district governments that represent the central government at the local level. Provincial governors are appointed by and report to the President, with executives that are

GHFRQFHQWUDWHGFDOOHG³GHOHJDWLRQV´RIOLQHPLQLVWULHVVXFKDVWKH0LQLVWU\RI+HDOWKRU(GXFDWLRQDW the local level. The central government replicates each of its agency functions at the provincial and district levels. District administrator appointments are made with input from provincial governors.

Unlike municipalities, which have some own-source taxation authorities, provincial and district government collect taxes and fees on behalf of the central government. Provinces are responsible for coordinating the activities of districts and responding to natural disasters whereas districts provide the same types of services (i.e., public illumination, cemetery maintenance) as autarquias.19

The focus of Mozambican decentralization, the 43 autarquias, are politically decentralized, with mayor-dominant governments advised by Municipal Assemblies both of which are elected to five-year mandates. The original 33 municipalities ZHUHFUHDWHGLQDQGFRQVLVWHGRIWKHFRXQWU\¶VPRVW populous cities, 23 provincial and district capitals, and 10 villages, with an additional 10 new

19 Provincial governors are advised by the newly elected Provincial Assemblies while district administrators appoint members of two consultative bodies: the Institutions of Participation and Community Consultation (IPPCs) and Consultative Councils. Although members are drawn from the FRPPXQLW\WKH,33&V¶SXUSRVHUHPDLQVXQFOHDU7KH&RQVXOWDWLYH&RXQFLOVDUHOLQNVEHWZHHQWKH district and the public and are charged with commenting on the District Strategic Plans and Budgets as well as monitoring the implementation of these plans. These Councils, unfortunately, can be unwieldy because they are comprised of between 30 to 50 individuals and their composition must include, among other requirements, persons appointed by the district administrator with input from the Chiefs of the Administrative Posts (Postos Administrativos), 30% must be public functionaries, 25% must be women. The combined sanctioning effect of both these bodies remains in question as the Consultative Councils are comprised of a large group of appointees from the district and administrative posts.

municipalities designated in 2008. In the three municipal elections held since 1998, FRELIMO has won all (or the majority) of mayoral seats and majorities in almost all Municipal Assemblies, combined with the results of national elections, makes Mozambique a de facto single-party state with limited political competition. However, decentralization did provide an opportunity for RENAMO, the main opposition party, and other parties to participate in politics. The ³GHFHQWUDOL]DWLRQSDFNDJH´and subsequent amendments assigned 29 competencies across nine areas which include health, education, sanitation, transportation and communication, and environmental protection and the authority to charge fees and taxes which include a municipal income tax, municipal property tax, economic activity tax, and market and trash fees. 20 Despite own-source revenue authorities, autarquias ,on average, receive at least 50% of their revenue from central government transfers.

The historical legacies of colonialism, socialism, destabilization, and decentralization affect governance in Mozambique and subsequently, government performance. These conditions have produced highly centralized municipalities, which are heavily dependent on the central government and donors. All municipalities are subject to these four conditions and thus, are held constant in my analysis. The following section addresses my methodology and justifies my case selection of four municipalities based on these factors²namely, participation, revenue and elections²that are hypothesized to influence municipal performance and vary across cases.

Method

Consistent with comparative casework done in the areas of behavioral economics and democracy development, this study employs an embedded single-case design using one case study country,

Mozambique (Yin 2003, 39-53). 21 As discussed, Mozambique provides a critical confluence of factors

20 Municipalities have two sources of own-source revenue: taxes (receitas fiscais) which include a municipal building tax, vehicle tax, municipal personal tax, and economic activity fee; and, fees (receitas não fiscais) which include property use, reservation of market space, parking, street vendor , parking, trash collection and advertising fees (Waty 2000). 21 The use of crucial cases is evident throughout comparative politics. Robert Putnam, Making Democracy Work: Civic Traditions in Modern Italy (Princeton: Princeton University Press, 1993). Graham T. Allison, Essence of Decision: Explaining the Cuban Missile Crisis, 1st ed. (Boston: Little

in its democracy development that provide insight into what factors affect service delivery that may be leveraged in understanding government performance in other, post-conflict, new democracies throughout the world. An embedded single-case design is perfectly suited for the study of performance at the local level because municipalities serve as the mini-cases and the services analyzed the units of analysis in each mini-case, thereby allowing for a replicable, comparative research design in various cities. This design is robust because it also permits for within-case control. Specifically, historical legacies and institutions, such as elections and budget-making rules, are held constant, because all mini- cases are chosen within one country with shared historical experience and institutions, such as the rules governing elections, staffing rules, and budgeting procedures. This permitted a closer examination of the variation of the dependent variable and allowed for a deterministic analysis of the degree to which the hypotheses are accurately reflecting variation in the dependent variable (Munck 2004).

Additionally, within-case analysis allows for the kind of process-tracing necessary to understand how various factors affect service delivery. Collier, Mahoney and Seawright state:

Within-case analysis proceeds by evaluating evidence about causal processes and mechanisms that link the independent variable to the dependent variable, looking for the specific ways that (in this case) civicness alters the goals and decision-making constraints of political, social and economic actors (Collier, Mahoney, & Seawright 2004, 96).

Within-case control is an important factor to consider in a research design because a dominant concern quantitative methodologists have of qualitative research is that the low number of cases chosen increases the likelihood of selection bias, therefore increasing the likelihood that the conclusions drawn are erroneously attributed to the explanatory variables selected rather than some other, unstudied, factor

(King, Keohane,& Verba 1994). The complex, relational nature of the dependent variable is suited to a

Brown, 1971). Ronald Rogowski, "How Inference in the Social (but Not the Physical) Sciences Neglects Theoretical Anomaly," in Rethinking Social Inquiry: Diverse Tools, Shared Standards, ed. Henry E; Collier Brady, David(Lanham, MD: Rowman & Littlefield Publishers, Inc., 2004). Deborah Bräutigam, "Institutions, Economic and Democratic Consolidation in Mauritius," Comparative Politics 30, no. 1 (Oct. 1997). Peter M. Lewis, Growing Apart: Oil, Politics, and Economic Change in Indonesia and Nigeria (Ann Arbor: University of Michigan Press, 2007). Guillermo A. O'Donnell, Modernization and Bureaucratic-Authoritarianism: Studies in South American Politics, vol. 9 (Berkeley: Institute of International Studies, University of California, 1973). The New Authoritarianism in Latin America, ed. David Collier (Princeton: Princeton University Press, 1979).

research design that permits examination of ³FDXVDOSURFHVVHVDQGPHFKDQLVPV´LQPRUHGHSWK$VD result, embedded single-case design, with multiple mini-cases for comparison of services on a cross-city basis, best allows for this type of analysis while mitigating for selection basis, using within-case control.

$VSUHYLRXVO\VWDWHGWKLVVWXG\¶VOHYHORIDQDO\VLVLVWKHPXQLFLSDOOHYHOEHFDXVHLWLVWKHORZHVW

OHYHORIYDULDWLRQRQERWKGHSHQGHQWDQGLQGHSHQGHQWYDULDEOHV%HFDXVHRI0R]DPELTXH¶VFHQWUDOL]HG system of government, the party elected nationally is in power at the district and provincial levels so the only level where political variation is possible is the municipal level. Furthermore, due to decentralization reforms in Mozambique, municipalities have authorities to raise revenue through a variety of fees and taxes, which are not available to the district or provincial level, making studying how own-source revenue capacities affect performance difficult.

Review of Factors for Selection of Municipal Cases

This project measured service delivery and three factors that theory tells us affect it (see Chapter 2) in four municipalities. In this section, I discuss how I selected these four cases. First, I gathered available secondary data for my independent variables (participation, revenue, and elections). While I was able to gather data on revenue and elections for all thirty-three cases, I only had participation data on twelve As a result, I examined how these variables worked in conjunction with each other to pick the cases based on the independent variables for which I had data (see Tables 1.1). I assigned a point value of Low (1), Medium (1.5) and High (2) for each independent variable. For participation, I examined secondary data on civil society. For revenue, I determined the per capita average for aid and all other sources of revenue and then assigned a rank depending on whether the per capita for a particular municipality was above or below the average. Lastly, for elections, I assigned a Low, Medium or High ranking contingent upon whether there was no change in party or mayoral candidate (L), no change in party but a change in mayor (M), change in party and change in mayor (H). I added these

point values and was able to choose four cases from these three categories. Those cases were Beira and

Dondo (High), Maputo (Medium), and Matola (Low) (see Tables 1.1 and 1.12).

I found some interesting patterns that I discuss below that led me to focus on examining my dependent variable²municipal performance²in four municipalities. In arraying my data, however, I found substantial gaps, as discussed, especially in data on participation. As a result, I validated my selection by conducting more than 60 field interviews with the World Bank-sponsored Municipal

Donors Working Group, Mozambican government officials from the Ministries of Planning and

Development, State Administration and Finance, the National Association of Municipalities of

Mozambique (ANAMM), NGOs and academics. Chapter 3 then empirically confirms that there is variation in my dependent variable²municipal performance²through multiple, first-hand observational experiments that took snapshots in time through fieldwork in 2008 and 2009. Then, I spend three subsequent chapters analyzing how my independent variables²civil society, revenue, and elections²affect municipal performance. Specifically, I conducted eleven focus groups where I asked residents to discuss municipal performance and 120 semi-structured interviews in four cities through field work in 2008 and 2009 and reviewed four years of budget data for all four cases.

I acknowledge that there are two critiques to this approach. First, positivist approaches posit that ideal case designs typically involve large-n studies (King, Keohane, and Verba, 1994). These scholars further that cases should be chosen based on variation of the independent variables to ensure that the

RXWFRPHEHLQJPHDVXUHGLVDWWULEXWDEOHWRWKHIDFWRUVSRVLWHGLQWKHK\SRWKHVLV³7KHEHVWµLQWHQWLRQDO¶

GHVLJQVHOHFWVREVHUYDWLRQVWRHQVXUHYDULDWLRQLQWKHH[SODQDWRU\YDULDEOH«ZLWKRXWUHJDUGWRWKHYDOXHV

RIWKHGHSHQGHQWYDULDEOHV´ Ibid, 140). The second critique of this approach is that four cases are not enough to explore the variance that three factors have on service delivery. King, Keohane and Verba

DUJXHWKDW³OLPLWLQJWKHQXPEHURIH[SODQDWRU\YDULDEOHVIRUZKLFK« RQH ZDQWVWRPDNHFDXVDO

LQIHUHQFHV«PD[LPL]H V OHYHUDJH´ .LQJ.HRKDQHDQG9HUED 7KHFKDOOHQJHP\DSSURDFK potentially poses is that changes in the dependent variable could be the result of something other than what I hypothesize in the relationship between my explanatory and dependent variables.

My view is that exploring relationships and processes between donors, the central government, municipalities, and residents required a small-n, case study approach. It provided the analytical leverage I need to see how multiple factors affect municipal performance. Furthermore, this approach allowed me to select cases based on both dependent and independent variables. While choosing cases based on explanatory variables, regardless of the values of dependent variables, may be the best

³LQWHQWLRQDO´GHVLJQ.LQJ.HRKDQHDQG9HUEDDFNQRZOHGJHWKDW³VHOHFWLRQVKRXOGDOlow for the

SRVVLELOLW\RIDWOHDVWVRPHYDULDWLRQRQWKHGHSHQGHQWYDULDEOH´ .LQJ.HRKDQHDQG9HUED, 129).

6LPLODUO\WKH\DFNQRZOHGJHWKDW³UDQGRPVHOHFWLRQLVQRWJHQHUDOO\DSSURSULDWHIRUVPDOO-QUHVHDUFK´ an approach this study adopts (King, Keohane, and Verba 1994, 128). As a result, I opted for a within- case research design to mitigate selection bias. Collier, Mahoney, and Seawright identify within-case

DQDO\VLVDVDZD\WR³PDLQWDLQ« WKH «IRFXVRQWKHRULJLQDOGHSHQGHQWYDULDEOH«E\HYDOXDting evidence about the causal processes and mechanisms that link the independent variable to the dependent

YDULDEOH´ %UDG\DQG&ROOLHU 2004, 93,96).

My choice of a qualitative, in-depth approach enables me to distinguish and appropriately attribute causal effects to the factors I posit affect municipal performance thus allowing me to determine the validity of my hypotheses. Specifically, an embedded single case design, permits me to: address multiple factors which affect performance, apply the principles of within-case control which allow me to examine specific causal relationships, and further control for these effects by having a replicable case design that permits me to compare outcomes across four cities (Yin 2003). According to Munck, within-FDVHFRQWUROGLVWLQJXLVKHV³EHWZHHQRXWFRPHVWKDWDUHV\VWHPDWLFZLWKUHVSHFWWRDJLYHQWKHRU\

DQGRXWFRPHVWKDWDUHUDQGRPZLWKUHVSHFWWRWKDWWKHRU\´ 0XQFN, 117). In so doing, I can compare outcomes and determine, through careful analysis of causal processes in each case, if the factors I posit affect service delivery do.

Explanatory Variables

My first explanatory variable was participation. I defined ³pDUWLFLSDWLRQ´as the engagement of groups of individuals, either formal or informal, that lobby or pressure the municipality to provide services. That pressure can come in the form of an association that facilitates interaction between the municipality and a group of individuals, such as Merchant Associations do for vendors. Alternatively, that pressure can be generated through participatory processes that connect the municipality with residents through regularly held hearings on topics of interest, as was the case with 'RQGR¶V participatory budget process. Or, pressure can, of course, come from the exercise of voice in the form

RISURWHVWVRUER\FRWWVRIVHUYLFHVDVZDVWKHFDVHLQ0DSXWR¶VVROLGZDVWHPDQDJHPHQWVHFWRU I also refer to this as civil society.

Civil society is not the private sector. However, as my evidence demonstrates, there were instances in this study in which civil society organizations became city contractors. As I discuss in

Chapters 2 and 5, in order to study how participation affects municipal performance, I included these organizations in my review because they are a part of civil society and provide insight of how participation can led to longer-term improvements in government performance. As I discuss in greater detail in Chapter 5, while there are some strong national civil society organizations, civil society in

Mozambique is weak for two reasons. First, public and private institutions are underdeveloped.

Second, there is a lack of ZKDW0DWWHVDQG6KHQJDFDOO³FULWLFDOFLWL]HQVKLS´RUinformed citizens that question their government. They state, ³WKHFRPELQDWLRQRI0R]DPELFDQV¶YHU\KLJKOHYHOVRIWUXVWLQ leaders and institutions with very low levels of commitment to democracy means that they present precisely the opposite archetype: WKDWRIµuncritical citizenship¶ ³ 0DWWHVDQG6KHQJD2007, 1).

Of my universe of thirty-three cases, I had limited data for twelve. I proceeded with the analysis of secondary material and verified my observations with field interviews. For the municipalities for which I had data, participation manifested itself either in informal self-help groups or, more broadly, the presence of a participatory process or regular channels of communication that are in place to facilitate citizen feedback to that municipality. Those cases where participation was high I found active

self-help groups or communication channels, formal or informal, between citizens and the town councilors or municipal staff. Cases in which I found low participation were characterized either by bureaucracies that limited or withheld information about municipal activities or where civil society organizations were not active. For the twelve cases for which I have data on participation, I am able to discern that five municipalities have high participation (Cuamba, Dondo, Maputo, Nampula, and

Quelimane) and seven had low participation (Beira, Chimoio, Gurue, Matola, Monapo and Nacala) (see

Table 1.1).

I then examined my second variable, which was resources, or revenue. I examined aid and both central government transfers and own-VRXUFHUHYHQXH FODVVLILHGDV³RWKHU´LQWKLVDQDO\VLVVHH7DEOH

1.1) and compared the average per capitas of each municipality against the averages for both aid and

³RWKHU´WRGHWHUPLQHKLJKDQGORZUDQNLQJV$FFRUGLQJWRWKH:RUOG%DQNH[FOXGLQJDLGWKHWRWDO budget for all municipalities was 1.1 billion MTn or USD$46 million in 2006, out of a state budget of over USD$3 billion, resulting in an average municipal per capita expenditure of USD$11.22

Municipalities received on average 42 percent of their revenue from the central government in 2007

(Anexo 2006-2007; Anexo 2008). Despite municipal dependence on the central government and low rates of spending, municipal own-source revenue grew, outpacing central government transfers in 2006 and 2007.23 The largest growth from 2007 to 2008 was in own-source receipts or taxes, which grew 27 percent and fees, which grew 19 percent from 2007 levels (Anexo 2008). Average municipal per capita

UHYHQXHIRU³RWKHU´VRXUFHV QDPHO\FHQWUDOJRYHUQPHQWWUDQVIHUVDQGRZQ-source revenue) for all 33

22 Author calculations. ConversiRQWRGROODUVGRQHE\DXWKRUXVLQJWKH&HQWUDO%DQNRI0R]DPELTXH¶V conversion rates dated 6/18/2009. 23 Own-source receipts have grown 25 percent from 2006 to 2007 in comparison to central government transfers, which have grown only 20 percent. Growth in receipts comes primarily from two types of receipts²current receipts and transfers. Current receipts are comprised of own-source revenue² primarily fees, like market and trash fees, and taxes, like personal and property taxes whereas transfers tend to include mostly the Municipal Compensation Fund (Fundo de Compensação Autárquica), which pay for non-elected municipal personnel. Current receipts have grown 68 percent since 2000, from $9.2 million to $19.2 million in 2007 (Anexo 2006-2007)(Anexo 2008). Calculations by author. http://www.mpd.gov.mz/orcamento/empresas/Anexo%20Autarquico%202007%20original.pdf and http://www.dno.gov.mz/docs/orc_estado/Proposta_OE_2008/Anexos_Informativos_ao_Orcamento/Aut arquias/Informa%C3%A7ao_Financeira_das_Autarquias.doc

municipalities from 2000 to 2008 was USD$6.40 from 2000 to 2008.24 Four municipalities had per capitas above the average (Beira, Dondo, Nampula and Vilankulos), with eight municipalities below the average (see Table 1.1).

While aid is also revenue, I examined it separately because Mozambique¶VEXGJHWUHIOHFWVDKLJK level of aid and I needed to isolate aid to compare the magnitude of difference between aid and other sources of revenue to understand the potential influence of aid on municipal performance. Two issues arise. First, if cases do not all have project-specific aid²especially in the sectors I plan to study, namely, trash collection and market management²then I could be biasing my selection by choosing a case that has no (or little) donor investment in a key area of study. Second, sector specific project assistance would improve services in a given sector, at least in the short-term.

I address these issues in two ways. First, I examine project data for aid to municipalities from

2004 to 2007 according to a comprehensive donor database called ODAMoz. According to this source, all but one municipality received project assistance for the time period studied (see Table 1.1).25 Once I narrowed the field of choice to four cases (as I will explain below), I was able to more closely examine the budgets of these cases. While different donors were in different municipalities and the levels of assistance varied, donors are present in all four municipalities (see Table 5.7 and discussion in Chapter

5 7KH:RUOG%DQN¶V3'0, which provided multi-sector assistance and included targeted assistance to both market improvements and garbage collection, was also available in all four municipalities. As

24 I selected cases that were in each category of municipal receipt on a per capita basis which reflect high and low values, using per capita U.S. dollars as the common unit of comparison. Municipalities are obligated by law to submit their financial statements to the Ministry of Finance. Article 24 of Law no 9/2002 of 12 of February, 2002 created the System of Financial Administration of the State (SISTAFE). Article 9 of Law 11/97 of 31 of May, 1997 the Law of Finances and Patrimony of the Local Autarchies require the central government to present a unified state budget to the National Assembly which includes financial information about the municipalities. Data for 2007 and 2008 where located in these annexes. The remaining data was gathered from the Conta Geral do Estado for respective years. 1999 and 2003 were not available. Data for 2008 is budgeted, not actual. There is enough variation across the five categories of revenue and expenditure²capital and current receipts, transfers and capital and current expenditures²to conduct this analysis. The project examined data from 2000 to 2008 for all 33 municipalities across these five categories of data to determine what patterns emerge. Specifically, municipalities were clustered in high and low categories based on average per capita receipts for each year. 25 I found no aid data for Manhiça.

such, I can defensibly argue that I hold sector-specific aid constant. However, there is variation in the levels and types of intervention for each of the donors present in the four cases I chose. Of particular note is the absence of a second-stage of funding in Matola for PDM. As I address in Chapter 5, the absence of this second tranche of funding is indicative of some concern over municipal capacity in

Matola that was echoed by at least one other donor. This finding underscores my conclusion of the importance of leadership in guiding municipal performance that I discuss in both Chapters 6 and 7.

Second, specific aid would improve services in the short-run because the municipality would benefit from a direct infusion of funding to a particular area. However, this short-term improvement in service would neither reveal much about the dynamics behind sustainable service improvements nor tell me how the hydraulics of resource and revenue pressure affect municipal performance. While sector- specific aid would understandably improve performance in that sector, my research found that service in the three cases was able to be sustained because 1) leadership (mayors, municipal teams) was in place to manage resources and 2) own-source revenue increased, which, combined with the appropriate leadership, could be utilized to sustain service improvements as was the case in Maputo with trash fees, for example.

Although the national government receives a commanding share of its budget in foreign assistance, aid constitutes a small share of municipal budgets. According to my review of a comprehensive donor database, Mozambique received USD$1.50 billion in foreign aid, which is roughly 56 percent of the total resources available in the 2007 State Budget.26 In order to exercise some

26 In general, donors have a memorandum of understanding with a line ministry to fund projects within municipalities so calculating municipal aid receipt presents some issues because this aid is aggregated at a central government level and the municipality often does not have the direct control over these project funds as they would have over their own-source receipts. However, one source of donor data is the Overseas Development Assistance in Mozambique (ODAMOZ), an effort started in January 2005 by the EC Delegation to harmonize donor contributions in Mozambique consisting today of 23 donors and 11 UN agencies. The analysis of this data suffered from three constraints: first, ODAMOZ does not currently classify projects by municipality so although municipalities are named as project recipients, funds are often awarded for projects in multiple municipalities and districts, making the estimate going to the municipality just that: an estimate. Secondly, there are other programs, which are hosted in ministries that have a component of municipal funding that may not be captured by ODAMOZ. Lastly, municipalities do not have direct and sole control over these funds as they would over own-source revenues. Because agreements are typically bilateral arrangements between the government of

degree of control over such large flows, the Mozambican government requested that donors provide aid in the form of budget support, as evidenced by the 71 percent of aid going to the national government, with the remaining 29 percent going to the provinces (Ibid). The largest share of aid flows was to general budget support (24 percent).27 Despite the proportion of aid in the national budget, aid constituted an average of only 9 percent of the municipal budget. 28 Therefore, although aid receipt is high in Mozambique and, arguably, municipalities receive more aid than is currently reported due to the fungability of foreign assistance, the central government retains the majority of control over both foreign assistance and transfers. Because of this dynamic, I used a donor database and estimated amounts to which municipalities have access to see if the four municipalities I chose had variation in

Mozambique and the donor country, municipalities usually must work with at least one line ministry in Maputo and the provincial government. Furthermore, funding is often irregular because of donor funding constraints and so this makes accessing and using these funds cumbersome and complex. However, donors directly input this data on a quarterly basis and indicate what municipalities, districts, or provinces are receiving this funding. I equally weight contributions across all recipients, a method which has been reviewed by the donor agencies representatives who work with this data. I shared with and incorporated comments from these donors on both the methodology and the results. Furthermore, I use actual, not projected data and verified the top relevant projects with actual site visits and interviews with donors. Lastly, it is the only available comprehensive database with this information. Understanding the limitations and benefits, this analysis still provides some useful insights.

The website for the database is www.odamoz.org.mz. Donors are: ADB, Austria, Belgium, Canada, Denmark, EC, Finland, Flanders, France, Germany, Ireland, Italy, Japan, MCC, Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, UK, USAID, and World Bank. UN agencies are: FAO, Global Fund, UNDP, UNESCO, UNFPA, UNHABITAT, UNHCR, UNICEF, UNIDO, WFP, WHO. The ODAMOZ database is a live, updated database in which donors regularly input disbursements and projections on a quarterly basis. As a result, this analysis represents only a snapshot of disbursements from December, 2007 to June, 2008. Although projections were available through 2011, actual disbursements available at the time of the analysis from 2004 to 2007. Mozambique Donor Atlas 2008, p. 10. Source: www.odamoz.org.mz, calculations by author. 2007 population estimate used. 27 Health (15 percent), transport and storage (10 percent), and government and civil society (8 percent) together with general budget support comprise over half of aids flows by sector in 2007. Five donors constitute almost 50 percent of the total aid flows to Mozambique in FY 2007. They are the World Bank ($239 million USD); the EC ($195 million USD); the UK ($111 million USD); USAID ($106 million USD); and Sweden ($98 million USD). 28 Nine percent is the flat average of 2006 donations, which is 12 percent and 2007, which is 7 percent calculated by the author. There is a rather large difference between donation to municipalities each year, which is consistent with the budget data for at least one city in the study and reflective of the general trend of donors in Mozambique. Aid regularly varies depending on the donor and the project. Figures from the Ministry of Planning and Development, Anexo 2007. Calculations by author.

this factor.29 Per capita average aid from 2004 to 2007 was USD$18.53 (see Table 1.1). Four municipalities were above the average (Beira, Dondo, Maputo and Nacala-Porto) while the remaining eight municipalities fell below the average.

The last factor for which I sought variation was on voting as a proxy for measuring how democracy compels locally-elected officials seeking election/-re-election to improve municipal services in the fulfillment of campaign promises. 30 While voting, by definition, varies because mayors across different cities will capture different numbers of votes, the patterns of party and mayoral turnover²the result of voting²vary in comparable ways over time. In the new political climate that decentralization provided,

I expected that the incumbents and challengers will compete for votes which will require the ability to not only meet the political demands of the party but also the ability to led and manage municipal staff, resources and relationships with relevant stakeholders which include not only the central government donors, as it did before but also voters. I am not implying that elections with turnover lead to high performance. Rather, I expected that elections with multiple parties competing and multiple candidates vying for the mayoral seat provide a higher incentive for incumbents and challengers to fulfill campaign promises. For example, an incumbent who has been re-elected and ran uncontested may be less inclined to fulfill campaign promises between mandates than an incumbent who faces several challengers.

As a result, I examined two factors when choosing cases: mayoral turnover and party turnover in the 1998 and 2003 municipal elections. If parties switched from 1998 to 2003, I ranked the election highly competitive (H). If the mayor switched but not the party, I ranked it medium (M). Lastly, if

29 Endogeneity is a concern here because municipalities could be awarded aid because they are high- performing and not the reverse, as I argue (aid awarded first and that improves performance). I address this in subsequent chapters by discussing the circumstances under which aid is awarded in each case as it relates to trash collection and vendor services. I trace how project-specific awards make discreet improvements in the performance of trash collection and vendor services over a five-year period. 30 Period under review in total is 1998-2008 because decentralization began in 1997, with the first municipal elections in 1998. For the purposes of analyzing electoral trends, data is available since 1998 and so the study reviews data for all three elections. However, for case selection purposes, only the changes between the 1999 and 2003 municipal elections were considered as the 2008 municipal elections had not yet taken place. The changes between the 2003 and 2008 elections will also be discussed as supporting the case selection choice after the fact. The period for review of budget data for all cities is 2004 through 2007 because the mandates for the mayors under direct review is the period from 2004 through 2008 (2008 budget data was not publicly available at the time of the research).

neither party nor mayor switched, then I ranked the election as not competitive or low (L) (see Table

1.1). The first municipal elections were held on June 30, 1998 amidst protests of administration issues, a

RENAMO boycott, and electoral fraud.31 As a result, mayoral turnover did not exist nor did rigorous party competition because RENAMO boycotted the election. FRELIMO candidates won all mayoral seats and FRELIMO dominated all the municipal assemblies. While these first elections required competition where it did not exist before, in the context of a new multi-party democracy, I ranked the

1998 elections as not competitive because only one party dominated (L) (see Table 1.1).

The 2003 municipal elections, held on November 19, 2003, saw more participation of political parties and therefore, more competitive elections.32 The results were still overwhelmingly in favor of the dominant FRELIMO party, with 28 municipalities electing FRELIMO mayors and 30 electing majority-

FRELIMO municipal assemblies. However, I ranked three elections competitive because these municipalities elected RENAMO mayors²Beira, Nampula, and Nacala-Porto.33 Two elections ranked medium because there was a change in FRELIMO mayor but not party²Maputo and Chimoio. The remaining seven were ranked as low because there was no change in mayor or party.

Based on this review, three municipalities ranked high with a score of 8²Beira, Dondo and

Nampula. Five municipalities were ranked as medium with a score in between 6 and 7.5²Maputo,

Nacala-Porto, Cuamba, Quelimane, and Vilankulos. The remaining four municipalities²Chimoio,

Gurue, Matola, and Monapo²ranked low (5 to 5.5 scores). I chose Beira and Dondo with high scores,

31 The RENAMO Electoral Union is often referred to just as RENAMO and is a coalition of 12 parties. RENAMO alleged problems with electoral law and fraud in voter registration, which led to a massive absence of opposition parties in the 1998 elections. Consequently, FRELIMO ran unopposed in 81 percent of municipal assembly races and 58 percent of mayoral races. However, reflective of general district and discomfort, voter turnout was extremely low (less than 15 percent) ("Observing the 2003 Elections" 2005, 20). 32 Revisions in 1999 and 2002 in electoral law led to better organized and regulated elections. Reforms included that the managing political body of the elections, the National Commission of Elections (CNE), was required to draw its 18 members from various political parties based on their strength in the National Parliament. There was also a clearer demarcation between the CNE¶VUROHZKLFKZDVWRVHW electoral policy within the parameters of the law and the Technical Secretariat of Electoral Administration (STAE), which would actually manage and administer the elections under CNE supervision. 33 The five municipalities that elected RENAMO mayors were Nampula, Angoche, Ilha de Mocambique, Nacala-Porto, and Beira. The three municipalities that elected the RENAMO-majority municipal assemblies were Angoche, Ilha de Mocambique and Nacala-Porto.

Maputo with a medium score, and Matola with a low score. Due to the lack of data, I independently verified my findings by conducting sixty semi-structured interviews with Mozambican government officials, NGOs, and donors who work in the area of decentralization in Mozambique. I asked open- ended questions regarding what municipalities might prove to be interesting based on my theory of participation, resources, and voting affecting municipal performance. My initial selection was corroborated by these interviews.

Dependent Variable: Municipal Performance34

From this starting point, I then examined municipal performance to confirm that variation existed on my dependent variable. I define and measure municipal performance as the provision of two services²trash collection and market management²for two reasons. First, under decentralization, municipal government has full authority to both provide and collect fees related to these two services.

As I am interested in understanding how both political and revenue pressure compel municipalities to improve service provision, I would expect to see how variations in lobbying and fee collection change over time to understand how these pressures affect municipal service delivery. Second, both services constitute a commanding share of own-source revenue. From 2004 to 2007, market and trash fees constituted on average 35 percent of own-source revenue, with market fees comprising the majority average share of own-source revenue (55 percent) for the same time period. Municipalities both have control over and depend on revenue from these two services and as such, would focus on the provision of these services and have the authority to make changes to improve them, making them the most appropriate services on which to focus.

I measured municipal performance using professional rankings and performance assessments by both independent, Mozambican-based non-profits, and the Ministry of State Administration, which oversees various aspects of municipal management, and interviews with the National Association of

Municipalities of Mozambique (ANAMM) and the Municipal Working Group, a group of donors

34 I use performance and service delivery interchangeably.

providing funding for decentralization based in Maputo. I found that Maputo, Beira and Dondo were high-performing and Matola, low-performing. According to Professional Management Review (PMR), a South-African ranking firm, Maputo and Beira were ranked as the two-highest performing municipalities in its 2007 and 2008 surveys (see Table 1.2) (PMR 2008). 35 Maputo was listed in all categories occupying one of the top four slots for attracting tourism, foreign investment, local

LQYHVWPHQWDQGMREFUHDWLRQ%HLUD¶VPD\RU David Mbepo Simango, was voted number one in mayors

GHVHUYLQJVSHFLDOUHFRJQLWLRQIRUDFKLHYHPHQWVRYHUWKHSDVW\HDUDQG0DSXWR¶VPD\RU'U(QHDV

Comiche, was ranked second. Dondo demonstrated good management with evidence of tangible results such as a new road or rehabilitation of a dump. Rankings and interviews suggested Matola as a low- performer due to a lack of security and trash collection. As a result of this review, I found variation in municipal performance in the four cases I chose: Maputo, Dondo and Beira for high performance and

Matola for low performance. I confirmed this initial review of municipal performance by conducting first-hand observational experiments to examine municipal performance in these four cases. I present these findings in Chapter 3.

Data Collection36

As previously stated, I employed an embedded single case-study approach using four cities as mini-cases to explore variance in municipal performance. One of the strengths of within-case design was that I was able to examine the effects of my independent variables on performance across cases. As

H[SHULPHQWV«WRIROORZD³UHSOLFDWLRQ´ORJLF

35 Also see PMR Vol. 20 Issue 03 which published the 2008-2009 results surveys, reiterating these findings. 36 Measurement is an issue with which both quantitative and qualitative research struggles. However, the study adopts a few WRROVWR³EULGJHWKHGLYLGH´DV6LGQH\7DUURZRIIHUVQDPHO\SURFHVVWUDFLQJ focus on tipping points, point of departure from quantitative data, and triangulation which are uniquely VXLWHGWRLQYHVWLJDWHSHUIRUPDQFH6LGQH\7DUURZ³%ULGJLQJWKH4XDQWLWative-4XDOLWDWLYH'LYLGH´ Rethinking Social Inquiry: Diverse Tools, Shared Standards, eds. Henry E. Brady and David Collier, Lanham, MD: Rowan & Littlefield Publishers, Inc., 2004: 174.

comparison in all four cases of data collected through the application of the same methods. In the

WUDGLWLRQRI5REHUW3XWQDP¶VUHJLRQDOVWXG\RI,WDO\,QRWRQO\XVHGPXOWLSOHTXDQWLWDWLYHDQGTXDOLWDWLYH data sources, I utilized experimenWVDVGLVFXVVHGLQWKH&KDSWHUWRWULDQJXODWHGDWDSRLQWVDQG³JDLQ

LQIHUHQWLDOOHYHUDJH´ 37 It is in this spirit that this study relies on four major sources of data collected from April 2008 to April 2009. First, I conducted semi-structured interviews of 120 government elites, donors, non-governmental organizations, Merchant Association vendors and Market Chiefs in 14 markets that represented three types of market²formal, informal and a mixture of formal and informal that are co-located. The specific questions varied depending on each group but I asked each group for their perspective on Mozambican decentralization and what specifically they think has affected municipal performance²both in general and specific to trash collection and vendor services. Second, I conducted eleven focus groups of 100 participants in formal and informal neighborhoods on trash collection services where I asked group members to describe the current state of services. Then, I asked them to identify the volume of trash in their neighborhoods and how much ability to pressure the municipality they felt they had to improve collection.38 Third, I examined municipal budgets, planning

37 3XWQDPFRQGXFWHG³FLWL]HQFRQWDFWH[SHULPHQWV´LQWZHQW\UHJLRQVIRr one month through a survey in which regional bureaucracies were monitored to see how typical requests by anonymous citizens for LQIRUPDWLRQZHUHKDQGOHG6LPLODUO\+HUQDQGR'H6RWR¶VMystery of Capital provides an excellent example of a study from whicKWKLVSURMHFWGUHZVRPHRILWV¶H[SHULPHQWDOPHWKRGV'H6RWR monitored in at least three countries²the Philippines, Egypt, and Haiti²the privatization of land by monitoring the application process for licenses, the steps it took, and various agencies that were involved, until property rights were granted. This project requested assistants to anonymously apply for ambulatory vendor licenses and track the progress in Maputo, Beira and Matola until the license was received. In all three cases, the licenses were received but the steps, costs, and length of time varied. Robert Putnam, Making Democracy Work: Civic Traditions in Modern Italy, Princeton: Princeton University Press, 1993:190. He conducted a series of institutional/political case studies in six regions where he analyzed the internal politics of regional institutions on political developments. He used local press as well as interviewed political leaders, party representatives, civil servants and interest group leaders. This is the case design on which this study is based. 38 For the focus groups, although I selected the two target neighborhoods, in all the municipalities, I had to rely either on the Neighborhood Secretary or Secretario do Bairro or the municipality itself to select the participants. Although this lead to selecting individuals who typically were a part of the neighborhood leadership, which could have led to bias in favor of the city in their responses, I felt that these individuals were the people who advocated for services for their communities and were extremely DZDUHRIWKHPXQLFLSDOLW\¶VZHDNQHVVHVZKHQLWFDPHWRSURYLGLQJWKHPVHUYLFHV)RUPXQLFLSDO management of markets, I interviewed the market leadership, which, although consisting of municipal employees, also consisted of vendor-elected Merchant Associations who are themselves vendors and consumers of the services the municipality is supposed to provide in the markets. To further mitigate

documents, studies and newspaper clippings since 1998. Lastly, I conducted experiments that involved the two municipal services under study²trash monitoring and applying for business licenses.

I conducted the majority of research for this project in-country in the months of September to

December, 2008²a period of time in which intensive campaigning for the third municipal elections held November 19, 2008²and again from February to May, 2009.39 I hired and trained eight

Mozambican researchers to facilitate the focus groups in three native languages (Xangana, Cena, and

Ndau) and Portuguese. Using performance measurement techniques from the public administration literature, I conducted focus groups using three forms²a scorecard, a rank list of what actions they would undertake to improve services, and four photos of trash accumulation²to augment the answers to the questions they gave in the discussion. They were translated into Portuguese and orally translated

against pro-municipal bias, I interviewed civil society organizations in each city, donors, and conducted two experiments to observe services first hand²trash monitoring and applying for an ambulatory business license. 39 One Mozambican researcher and friend cautioned against conducting the research at this time from a practical and theoretical standpoint. From a practical point of view, municipal staff and mayors with whom this project would have sought to interview would simply not be available and would not want to WDONWRDIRUHLJQUHVHDUFKHUDERXWWKHPD\RU¶VSHUIRUPDQFHGXULQJ a time in which he is trying to get re- elected. The campaign season prior to the municipal elections was two weeks and during this period individuals simply are not at work because municipal staff are campaigning as if they were private citizens. Indeed, as I observed firsthand, the culture is that everyone should be out campaigning during this period to re-elect the incumbent, which, more often than not, is the FRELIMO candidate. There is DQREYLRXVFRQQHFWLRQEHWZHHQPXQLFLSDOVWDIIV¶MREVHFXULW\DQG that of the incumbent that transcends party, however. Having observed campaigning in Beira, the main opposition city, other parties (and municipal functionaries) also campaign heavily during this period. Having conducted the research in three cities during this time, I found, with a few modifications, and a return trip to the fourth city after the elections that the research ran smoothly. Secondly, from a theoretical point of view, the data being gathered at this time could have arguably been skewed because elections occur every five years and municipal staff are going to want to portray the mayor in the best light. Furthermore, answers might be skewed because residents will hesitate to answer questions about mayoral performance for fear of retribution. Although it is true that this period might affect the answers I receive in interviews because of political considerations, , this project argues that the best time to study accountability is precisely before elections because it is when the mayor has most at stake. Furthermore, citizens are most aware of municipal performance (or lack thereof) and their exercise of the sanctioning power of their vote right before an election. Therefore, this project gathered robust data that might not have been available during another period of time. Lastly, explaining thoroughly the use of the data, benefits and cautions of participating in focus group discussion about the research topic and facilitating focus groups with native speakers of both Portuguese and the indigenous languages of each region (Xangana in Maputo and Matola and Ndau and Cena in Beira and Dondo), focus group participants were extremely open about their expectations and more than willing to share their thoughts on how their services could be improved.

into the local languages.40 For the focus groups, although I selected the two target neighborhoods, in all the municipalities, I had to rely either on the Neighborhood Secretary or Secretario do Bairo or the municipality itself to select the participants.41 Although this led to selecting individuals who typically were a part of the neighborhood leadership, which could have led to bias in favor of the city in their responses, I felt that these individuals were the people who advocated for services for their communities

DQGZHUHH[WUHPHO\DZDUHRIWKHPXQLFLSDOLW\¶VZHDNQHVVHVZKHQLWFDPHWRSURYLGLQJWKHPVHUYLFHV

For municipal management of markets, I interviewed the market leadership who, although consisting of municipal employees, also consisted of vendor-elected Merchant Associations who were themselves vendors and consumers of municipally-provided services. To further mitigate against pro- municipal bias, I interviewed civil society organizations in each city, donors, and conducted two experiments to observe services first hand²trash monitoring and applying for an ambulatory business license.42 I conducted all semi-structured interviews in Portuguese.

Plan for Dissertation

The purpose of this chapter was to lay the groundwork for this dissertation. My hypothesis is that decentralization, in theory, is supposed to promote accountability at the local level and hence, improve municipal performance, by encouraging public participation, resource generation, and political competition. In other words, decentralization should improve performance by increasing accountability

40 6HH&KDSWHU³:KDW0HWKRGVRI'DWD*DWKHULQJ6KRXOG%H8VHG"´LQ+DUU\+DWU\¶VPerformance Measurement: Getting Results, 2nd edition, Washington, D.C.: The Urban Institute Press, 83-117. 41 The two neighborhoods were chosen within the cement part of the city because the service, in all cities, did not regularly extend to the unpaved areas of the city, because of access issues. This presented a problem because all citizens who pay should receive service and all residents of the city have a right to access this municipal service. Many households outside the cement part of the city do not have electricity, and as such, do not pay for trash collection. To not exclude even the population that does not pay but still has a right to access this service (and do use it if only when they leave their neighborhoods to enter into the cement city to work), the project interviewed available community neighborhood associations who work in the area of trash collection, sanitation and public health to understand what, if any service, these neighborhoods get. I also discussed this issue with the municipal town councilors responsible for providing this service. 42 ,QNHHSLQJZLWK3XWQDP¶VFRUUHFWDSSOLFDWLRQRIWKHDQRQ\PLW\SULQFLSOH'RQGRZDVH[FOXGHGIURP this experiment as the research assistants were known to the individual who would be reviewing the license application.

between civil society, voters, and fee payers and elected officials by incentivizing them with political

(votes, lobbies) and revenue (fees and taxes) pressure (see Figure 1.1). This chapter addressed why this study is significant and necessary, how it will contribute to the larger discussion on performance in the context of decentralization, and why Mozambique is a natural laboratory in which to conduct this inquiry. I also discussed how the single case approach using an embedded case design allowed me to hold institutions²such as colonialism and decentralization²constant and explore how various factors affect performance.

Chapter 2 provides the theoretical framework on which this study is built. Consistent with the factors used here, theory tells us that performance is affected by participation (civil society), resources

(revenues), and politics (votes). The following chapter addresses in greater depth what scholars have said about these relationships and how my research contributes to the debates. Chapter 3 empirically establishes variance in my dependent variable²performance²by presenting findings from two observational experiments I conducted in the four case study municipalities. I present my evidence in chapters 4 through 6 that examine how each of the independent variables (participation, resources and political competition) affects performance. I conclude with a review and discussion of the implications of my findings in Chapter 7.

Figure 1.1. Hypothesis

Decentralization Accountability High Performance

Revenue Pressure (fees and taxes)

Political Pressure (votes, lobbies, and citizen participation)

Table 1.1. Factors used to select municipal cases

Civil No. Municipality Province Society Revenue (USD, per capita) Voting Total Aid (average: Other (average: Rank $18.53, 2004-2007) $6.40, 2000-2008) Rank 1998 2003 Rank 1 Beira Sofala L 32.46 6.57 HH L H LH 8 High 2 Dondo Sofala H 28.35 9.54 HH L L LL 8 3 Nampula Nampula H 11.89 10.23 LH L H LH 8 4 Maputo Maputo H 18.74 6.32 LH L M LM 7.5 Moderate 5 Nacala-Porto Nampula L 24.01 4.70 LH L H LH 7 6 Cuamba Niassa H 1.26 5.26 LL L L LM 6.5 7 Quelimane Zambezia H 3.94 5.93 LL L L LL 6 8 Vilankulos Inhambane L 3.32 7.96 LH L L LL 6 9 Chimoio Manica L 9.98 4.19 LL L M LM 5.5 Low 10 Gurue Zambezia L 0.00 1.59 LL L L LL 5 11 Matola Maputo L 3.18 2.42 LL L L LL 5 12 Monapo Nampula L 0.25 6.08 LL L L LL 5 H=2, M=1.5, L=1

Table 1.2. Municipal Performance and Explanatory Variables for Selected Cases

Case Performance Participation Revenue* (USD, per Mayor/Party capita) ¶WR¶ mandate)

Aid All other ($18.53) ($6.40)

Maputo Voted as best managed High Medium: Medium: FRELIMO : municipality in 2007 and 2008 $18.74 $6.32 Mayor by South African-based switched, party ranking magazine. same.

Beira Voted as one of the best Low High: Medium: RENAMO: managed cities in 2007 and $32.46 $6.57 Mayor 2008 by South African-based switched, party ranking magazine. Members switched. of public indicated that they are satisfied with mayor.

Dondo National and international High High: High: $9.54 FRELIMO: model of participatory $28.35 Same mayor, budgeting. Performance plans same party. indicate planned projects are completed in time.

Matola Citizens unsatisfied with Low Low: Low: $2.42 FRELIMO: PXQLFLSDOLW\¶VLQDELOLW\WR $3.18 Same mayor, provide security and trash same party. collection.

Note: Aid averages are from 2004 to 2007 and all other, which is constituted of own-source revenue and central government transfers), are from 2000 to 2008.

II. GROUNDING EMPIRICS IN THEORY

Decentralization has, not only an administrative value, but also a civic dimension, since it increases the opportunities for citizens to take interest in public affairs; it makes them get accustomed to using freedom. And from the accumulation of these local, active, persnickety freedoms, is born the most efficient counterweight against the claims of the central government, even if it were supported by an impersonal, collective will. Alexis de Toqueville (Bonnal n.d.)

Real decentralization is very limited in Mozambique. Dr. John Barnes, Maputo, Mozambique ("Decentralização em Moçambique" n.d.)

Introduction

The purpose of this chapter is to discuss the theory that led me to identify participation, resources, and elections as the three factors that affect municipal performance. Based on this discussion, I hypothesized that decentralization improves municipal performance by promoting accountability.

Specifically, decentralization gives civil society, voters, and fee payers an opportunity to hold local officials accountable by incentivizing them with political (votes, lobbies) and revenue (fees and taxes) pressure. I engage the decentralization, democracy and fiscal contract literatures to identify how participation, resources, and elections affect municipal performance. In subsequent chapters, I empirically explore how these factors affect municipal service in four Mozambican municipalities and find that municipal performance improves in ways that theory might not have predicted.

In addition to those factors, another critical factor²mayoral leadership²surfaced in my field research. The discourse of leadership within the context of democratization in sub-Saharan Africa tends to revolve around the predatory, tribal nature of African leaders, providing few examples of those successful leaders and what specifically makes them so. My research provides empirical evidence that leadership positively impacts performance because it is the intervening factor, which makes up for weak participation and low levels of both resources and control over those resources. In the next section, I outline my hypothesis. Then, using the decentralization literature as a launching point, I identify the three major accountability factors that affect performance. From there, I analyze how the scholarship in

each area treats performance. Then, I identify how my research contributes to these discussions. In the final section, I link the theory to my cases, providing the foundation on which the following three chapters rest. Then, in my concluding chapter, I make final observations and extend my findings to other areas of future study.

The Hypothesis

I hypothesized that decentralization improves municipal performance by promoting accountability.

Specifically, decentralization creates incentives, which motivate municipal government officials to act and opportunities for residents to pressure them in the following ways (see Table 2.1):

1. Resource pressure over local officials increases as new authorities put resources under the

control of local decision-makers; and,

2. Political pressure over local officials increases through new opportunities for the expression of

citizen preferences through voting and civil society organizations in municipal policy-making.

Decentralization is thought to improve municipal performance because it introduces institutional reforms²like local elections and own-source revenue authorities²which re-focus municipal accountability from the central government to residents. Whereas, prior to decentralization, local officials were appointed by, and therefore accountable to, the central government, now they are under pressure to seek election based on a platform of performance for residents. Similarly, decentralization can shift more ownership over resources to local officials by providing municipalities the authority to generate and keep more taxes and fees. Elections (specifically voting) and own-source revenue authorities (specifically for taxes and fees) provide political and revenue pressure residents can now use to motivate municipal performance because local officials are incentivized to earn votes and increase revenue over which they now have some control and residents, not the central government, can sanction

(or reward) municipal performance by withholding votes and revenue. Furthermore, by subjecting local

officials to the pressures of earning both votes and revenue, civil society also develops leverage because it can, in theory, now galvanize voters and tax payers to sanction or reward public officials.

Decentralization and Performance

Decentralization is all things to all people. 43 Cheema defines democratic governance as characterized by rather amorphous principles, such as accountability, participation and transparency,

ZLWKGHFHQWUDOL]DWLRQEHLQJ³FRQGXFLYHWRWKHDFKLHYHPHQWRIWKHVHSULQFLSOHV´ &KHHPD

.DGPLHO:HNZHWHREVHUYHGWKDW³GHFHQWUDOL]ation is viewed in many countries as a vehicle to promote participatory decision-PDNLQJORFDOLQIUDVWUXFWXUHGHYHORSPHQWDQGVHUYLFHGHOLYHU\´ :HNZHWH

242). One point of common agreement, however, is that the performance of decentralized regimes is largely determined by local government performance and, as such, decentralization scholarship serves as an appropriate springboard for my exploration of municipal performance. As decentralization scholars Cheema and Rondinelli acknowledged in their introduction, "one of the most critical issues in implementing decentralization...is identifying those factors that facilitate strong local government performance" (Cheema and Rondinelli 2007, 5-6). Crook and Manor identify administrative and revenue viability, autonomy, elections and the quality of interaction between elected officials, bureaucrats and the public as factors that affect government performance (Crook and Manor 1998).

Empirical research validates these factors. Civil society, technical capacity, regime support, and control

43 Decentralization is a transfer of administrative, fiscal or political authority away from the central authority to lower levels in a territorial hierarchy, which is generally classified in two ways: a devolution of power, which cedes authority to lower levels of government in such a way as to define autonomy of these units separate from the central government and deconcentration of power, in which central government powers are devolved to local representatives of the central government. Definitions adapted from Richard C. Crook and James Manor, Democracy and Decentralization in South Asia and West Africa: Participation, Accountability and Performance, Cambridge: Cambridge University Press, 1998:6-DQG$UXQ$JUDZDODQG-HVVH5LERW³$FFRXQWDELOLW\LQ'HFHQWUDOL]DWLRQ$)UDPHZRUNZLWK 6RXWK$VLDQDQG:HVW$IULFDQ&DVHV´Journal of Developing Areas, No. 33, Summer 1999: 473-502. )RUDFRPSUHKHQVLYHRYHUYLHZRIGHFHQWUDOL]DWLRQVHH*6KDEELU&KHHPDDQG'HQQLV$5RQGLQHOOL¶V Decentralizing Governance: Emerging Concepts and Practices, Washington, D.C. : Brookings Institution, 2007. Mozambique has undertaken all three types of decentralization, although fiscal decentralization has been considered by far the least successful. See Fernanda Faria and Ana Chicava, ³'HVFHQWUDOL]DomRHFRRSHUDomRGHVFHQWUDOL]DGDHP0RoDPELTXH´ZLWKILQDQFLDOVXSSRUWIURPWhe European Union, October, 1999. See: http://www.ilo.org/public/portugue/region/eurpro/lisbon/pdf/descentra_mocambique.pdf 50

over budget resources all contributed to decentralization in the Philippines (Guess 2005). An active civil society made local government more responsive in Bolivia (Faguet 2009). One study of 30

Mexican municipalities examined how political competition, elected officials and municipal staff capacities, and civil society participation affected municipal performance. It demonstrated that elections and civil society played a substantial role in fostering citizen awareness and political competition (Grindle 2007, 56-74). Elections empowered mayors in Bolivia, Uganda, and Brazil to improve municipal performance by requiring professional certification requirements for building standards and instigating environmental, waste management and social service programs (Cheema 2007,

176-177).

Based on this review, I explore how political pressure, specifically civil society and elections, and resource pressure, specifically, revenue, affect municipal performance in the areas of markets and trash collection in Mozambique (see Table 2.2). I posit that municipal performance will improve if the following three conditions are met. First, a strong civil society should exist to pressure a more responsive government to deliver better services. Second, municipalities would be more responsive to taxpayers if they rely on own-source revenue. Third, competitive elections incentivize mayors to improve municipal services to constituents. The following three sections root these expectations in theory.

Participation and Civil Society: External to and Extension of the State

Civil society as a concept is unwieldy and often efforts at clarification only serve to further obscure. The common element of these definitions is the manifestation of citizen interaction with the state. More specifically, in this study, I focus on how effective organizations and processes²various aspects of participation²were at pressuring the municipality to improve service delivery. I found an informal vendors lobby in Maputo to successfully lobby the CMM for favorable concessions in

municipal policy but WKHVHFRQFHVVLRQVGLGQRWPDNHYHQGRU¶VOLYHVHDVLHU,Q0DSXWRFLWL]HQV successfully protested the introduction of a flat trash fee in 2002, causing the CMM to work with a donor to rationalize the fee schedule and make it more fair, based on electricity usage. I also found in

Maputo that neighborhood organizations, with the help of donors, were able to build the capacity to lobby for and successfully win and implement municipal contracts to collet trash in informal neighborhoods. In Dondo, a participatory budget process instituted another mechanism by which citizens could work with the CMD to prioritize infrastructure investments. And, while Merchant

Associations played important roles in Maputo, Beira, and Dondo in facilitating municipal management of the markets, only in Maputo, do these organizations have access to funds to decide and implement changes vendors want to see. In sum, I found that participation in its various forms in these cases created links that informed municipal decisions and, in some cases, led to improvements in municipal performance.

In the decentralized context of Mozambique, a post-conflict, post-socialist new democracy, civil society can play many roles. In a new democracy, civil society in Mozambique can emerge as Larry

'LDPRQG¶V³UHDOPRIRUJDQL]HGVRFLDOOLIHWKDWLVYROXQWDU\VHOI-generating, at least partially self-

VXSSRUWLQJ´ZKHUHFLYLOVRFLHW\DSSHDUVDVOREE\LQJWKHVWDWHIRULQFUHDVHGDFFHVVWRUHVRXUFHVDQG services. thereby increasing voice and participation by demanding a share of resources and say in being governed. To that end, I would expect civil society to improve municipal performance because it applies pressure to the municipality to redistribute resources in the form of improved access to services.

With socialist and colonial roots, civil society in Mozambique also appears as an extension of the state, as exemplified in the Organization of Mozambican Women or Merchant Associations that may not be either voluntary or self-supporting, as described above, but charged with maintaining order rather than increasing access to public resources. Here, civil society primarily preserves the status quo, retaining public resources in the hands of a few elites. I would expect municipal services to suffer both because there is an absence of effective pressure groups and power is concentrated, potentially resulting in weak accountability.

The concept of civil society and how it interacts with the state is historically rooted in

Enlightenment Europe, where kings were increasingly pressured to serve an emerging state based on

UXOHRIODZ³,IWKHVWDWHZHUHWRHPERG\XQLYHUVDOLW\WKHQFLYLOVRFLHW\PXVWEHVHHQDVFUHDWLQJWKH

EDVLVRIFRPPRQZHOIDUHRXWRIWKHSXUVXLWRISDUWLFXODULQWHUHVWV´ &R[ More to the point, civil society was envisioned as being able to coalesce the common good from the pursuit of individual interests through the rise of the bourgeoisie (Ibid). Hegel, and later Adam Smith, saw civil society linked to market place incentivHVZKHUHLQGLYLGXDOVSXUVXHGLQWHUHVWVJXLGHGE\WKH³LQYLVLEOHKDQG´

%HFNHU 7RTXHYLOOH¶VVWXG\RIYROXQWDU\DVVRFLDWLRQVLQ$PHULFDDGGHGDGHILQLQJFKDUDFWHULVWLF to this concept²one in which civil society consisted of voluntary citizen participation in organizations.

Tocqueville, along with John Stuart Mill, and John Locke, described civil society as separate from state, where certain individual freedoms, such as freedom of speech and freedom of religion, are protected as rights of citizenship. In this ontology, civil society can negotiate with the state on the protection of these freedoms as rights of citizenship, in which the individual acknowledges the legitimacy of the state

DQGLWVH[HUFLVHRIFRQWUROLQUHWXUQIRUWKHVWDWH¶VDFNQRZOHGJPent of individual rights through the

FRPSDFWRIFLWL]HQVKLS%RWK/RFNHDQG5RXVVHDXGLVFXVVWKLVDVWKH³VRFLDOFRQWUDFW´EHWZHHQWKHVWDWH and its citizens.44 As a result, civil society as an independent force has become a seemingly universally-associated precondition to democratic development and consolidation because of the reciprocal and accountable nature of the state-citizen relationship.

In the post-Cold War era, some scholars saw civil society as a way to democratize former

Communist states, deYHORSLQJYROXQWDU\DVVRFLDWLRQVIUHHVSHHFKDQGSROLWLFDOFRPSHWLWLRQ³7KHWHUPV

RIFLYLOVRFLHW\LWVDWWUDFWLYHFRPELQDWLRQRIGHPRFUDWLFSOXUDOLVP«PDNHLWDSSHDUKRSHIXOWRVRFLHWLHV

VHHNLQJWRUHFRYHUIURPWKHH[FHVVHVRIVWDWHVRFLDOLVP´ .XPDU3, 375). According to the

44 Rousseau stated WKDW³DOOOHJLWLPDWHJRYHUQPHQWLVµUHSXEOLFDQ¶´, meaning it derives its authority and legitimacy from its people (Book II). Furthermore, the importance of rule of law in a democracy and its connection to civil society can be traced to Rousseau. Law, in both the Weberian and Rousseaun views, are socially constructed and derive their legitimacy from the people subject to them. See Jean-Jacques Rousseau, The Social Contract, trans. Maurice Cranston, (Penguin Books: New York, 1968) and Max Weber, Theory of Social and Economic Organization, ed. Talcott Parsons, (Free Press: New York, 1964). 53

democratization literature, as previously stated, ³FLYLOVRFLHW\LVWKHUHDOPRIRUJDQL]HGVRFLDOOLIHWKDWLV voluntary, self-generating, at least partially self-supporting, autonomous from the state and bound by a legal order oUVHWRIVKDUHGUXOHV´ 'LDPRQG 5REHUW3XWQDPDQG/DUU\'LDPRQGERWKEXLOWRQ this tradition of civil society in the context of democracy development and democratic consolidation.45

In this view, civil society organizations are individually-cohesive units and financially independent such that they can counterbalance the state. Diamond found that a critical component of civil society is

SHRSOH¶VULJKWVWRIUHHDQGRSHQVSHHFK/LNH7RFTXHYLOOH'LDPRQGFLWHVFLYLOVRFLHW\DVDEXOZDUN against authRULWDULDQLVPLQ$IULFDQVWDWHVVXFKDV1LJHULDDQG=DLUHEHFDXVHLWSUHVHUYHV³VRPHNHUQHORI

FLYLFQHVVLQFXOWXUHVRPHVHHGVRIKRQHVW\WUXVWVROLGDULW\HIILFDF\DQGKRSH´ 'LDPRQG $VLQ

3XWQDP¶V,WDO\FLYLOVRFLHW\LVQRWMXVWDVSDFHIRUIUHH VSHHFKEXWD³FLYLFFRPPXQLW\´WKDWIDFLOLWDWHG the development of democracy in and economic development of Northern Italy (Putnam 1993).

6SHFLILFDOO\3XWQDP¶VUHVHDUFKLQ,WDO\KLJKOLJKWHGDFRPPRQOLQNIRXQGDPRQJth century political philosophers, ZKLFKLVWKHFRQFHSWRI³VRFLDOFDSLWDO´0RGHUQ-day rational choice approaches to solving collective action problems of society relied heavily on individuals choosing to cooperate because it was in their best interests. Putnam found in Northern Italy that enforcing cooperation occurred through dense networks of associations in which individuals collaborated because they had a past built on trust. Some argue that this idea is culturally and chronologically rooted in Enlightenment

Europe and therefore, not particularly applicable in modern-day sub-Saharan African development because it excludes kinship networks (Woods 1992, 77-100).46 As a new democracy, this definition of civil society provided a useful heuristic for examining civil society organizations in the four case study municipalities in Mozambique because, as decentralized entities, the theory is that democratic openings encouraged associations to participate in public life and pressure these governments to provide better

45 In Polyarchy, Dahl argues that the pluralistic governance of polyarchy is rooted in its ability to foster participation and tolerate and promote opposition. From this perspective, Dahl describes the role of the political activist as an advocate for change. He argues that their role, like that of civil society under GLVFXVVLRQKHUHLV³FRPSHWLWLYH-FRRSHUDWLYH´ 46 Flavio Mulando gives a brief overview of the post-colonial state/civil society dynamic in his paper, ³23DSHOGDV2UJDQL]Do}HVGD6RFLHGDGH&LYLOQD)RUPXODomRGH3ROtWLFas Públicas em Moçambique: &DVRGR*HR3DUSD,,´&RQIHUHQFH3DSHUQRInaugural Conference of the Institute for Social and Economic Studies (IESE) Challenges for Social and Economic Investigation in Mozambique, September 19, 2007. 54

services. Here, as theory suggests, I would expect civil society to be an independent force, separate from the state.

However, Mozambique is a post-socialist, post-colonial state dominated by the same political party for over thirty years. As such, civil society that emerged organically here were not limited to voluntary associations that pressure the state but include mass organizations, such as the National Organization of

Mozambican Women, labor associations, and neighborhood cells organized to, exert state control.

From a theoretical perspective, Marx was skeptical of the idea of a common good emerging from individual interests. He saw civil society as a force of conflict, control and change through the unification of the proliteriat (Becker, 1932). Later, Gramsci, based on his observations of his native fascist Italy, built on this perspective of civil society being the space of hegemony and counter- hegemony. ,Q*UDPVFL¶VYLHZFLYLOVRFLHW\³KHOSHGFUHDWHLQSHRSOHFHUWDLQPRGHVRIEHKDYLRUDQG expectations consistent with WKHKHJHPRQLFVRFLDORUGHU´ (Cox 1996, 126). Janos Kornai examined

PDVVRUJDQL]DWLRQVDV³WUDQVPLVVLRQEHOW´RUJDQL]DWLRQV, which transmitted party ideology to the masses and kept order in neighborhoods and markets (Komai 1992).47 In evaluating the post-colonial Africa

VWDWH&KDEDODQG'DOR]UHLWHUDWHGWKLVYLHZRIFLYLOVRFLHW\WKDWLW³LVWKHFRQVWDQWLQWHUSHQHWUDWLRQRU

VWUDGGOLQJRIWKHRQH FLYLOVRFLHW\ E\WKHRWKHU WKHVWDWH ´ &KDEDODQG'DOR] 7KHVHJURXSVDV

³FLYLOVRFLHW\´DUHZHDN in that they typically do not challenge the state or assert an independent voice because they are a part of a state control mechanism +RZHYHUWKH\FDQDQGGRPDLQWDLQD³KHJHPRQLF

VRFLDORUGHU´48 However, as part of this order, these organizations do participate in governance, albeit in a more controlled fashion. Here, I would expect civil society to be part of the state.

47 The point about mass organizations is particularly salient here because Mozambique under Samora Machel pursued a command economy model of development, which extended to the creation of socialist social structures like block chiefs, neighborhood secretaries, provincial secretaries, and merchant associations to keep the ideological discipline of the dominant FRELIMO party. 48 An example of how civil society can maintain order where the state cannot and even provide services FDQEHIRXQGLQ(QULTXH'HVPRQG$ULDV¶DUWLFOHon Brazilian favelas entitlHG³)DLWKLQ2XU1HLJKERUV 1HWZRUNVDQG6RFLDO2UGHULQ7KUHH%UD]LOLDQ)DYHODV´Latin American Politics and Society, Vol. 46, No. 1, Spring 2004: 1-38. 55

A key difference between these two perspectives is that in the first, civil society is linked to autonomous, voluntary³DWOHDVWpartially self-VXSSRUWLQJ´ associations outside the state while in the second, civil society can be the dominated by or an extension of the state without access or control to resources. In Mozambique, I observed donor-funded neighborhood trash associations and an informal vendors lobby as separate from the state. Civil society in Mozambique includes organizations that

³SURYLGHµWUHQFKHV¶DQGSURWHFWWKHµIRUWUHVV¶RIVWDWHSRZHU´ 3DWQDLN &KD]DQVXPPDUL]HV this view of Africa civil society aV³DQRUJDQL]DWLRQ WKDW PXVWVLPXOWDQHRXVO\FRQWDLQVWDWHSRZHUDQG

OHJLWLPDWHVWDWHDXWKRULW\´ )DWWRQIRRWQRWH In Mozambique, I saw Merchant Associations as these extensions of the state. Civil society in this view can include organizations that might

RWKHUZLVHEHH[FOXGHGIURP'LDPRQG¶V³UHDOPRIRUJDQL]HGVRFLDOOLIHWKDWLVYROXQWDU\VHOI-generating, at least partially self-VXSSRUWLQJ´VXFKDVNLQVKLSJURXSVWKDWVHUYHGDVDQH[WHQVLRQRIWKHORFDOFRORQLDO government and, eventually, decentralized municipalities (Diamond 1999). Africanist scholars further observed that civil society is weak due to patrimonial tendencies of African states, which make independent resources for civil society scarce (Lewis 1992, Gyimah-Boadi 1996). While mass mobilization of local interest groups might provide enough resistance to central government pressures,

LQWKH$IULFDQFRQWH[W³HQGRJHQRXVSUHVVXUHJURXS V DUH QRWXSWRWKHWDVN´ $ZRUWL -RKQ

Makumbe cites self-sustainability for African civil society to be a major problem, which then requires reliance on foreign aid (Makumbe 1998). Within this context, the informal world of associations, kinship groups and self-help networks, groups that scholars like Larry Diamond might not include as

FLYLOVRFLHW\RSHUDWHLQZKDW5LFKDUG)DWWRQFDOOVWKH³VXEWHUUDQHDQFUDFNVRIW\UDQQLFDOUXOH´49

While not incompatible, as both democratization scholars and Marxist scholars rely on common foundations of Enlightenment scholarship that depicts civil society as coalescing mutual interests interacting with the state, these scholars have pivoted into distinct directions. If I relied solely on one

49 /DUU\'LDPRQGGHILQHVFLYLOVRFLHW\DVH[FOXGLQJ³LQGLYLGXDODQGIDPLO\OLIH´/DUU\'LDPRQG³ &LYLO6RFLHW\DQGWKH'HYHORSPHQWRI'HPRFUDF\´Estudio/Working Paper 1997/101, June 1997: 6, accessed at http://www.march.es/ceacs/ingles/publicaciones/working/archivos/1997_101.pdf, on 2/22/2011, see 5REHUW)DWWRQ³$IULFDLQWKH$JHRI'HPRFUDWL]DWLRQ&LYLF/LPLWDWLRQVRI&LYLO 6RFLHW\´African Studies Review, Volume 38, No. 2, September 1995: 69. 56

view of how civil society increases participation and voice²either as self-sufficient, voluntary organizations or protests lobbying the state or extensions of the state as associations linked to a socialist order²I would not adequately capture civil society in Mozambique and how it affects municipal performance. Specifically, I would have not analyzed how Merchant Associations act as interlocutors between vendors and municipal government to improve services to vendors because they are extensions of the state. Similarly, I would have disregarded those neighborhood trash organizations in Maputo that eventually became municipal contractors because they were no longer part of civil society but rather part of the private sector. As these examples demonstrated, I would, in essence, be omitting from my analysis processes or groups that might be defined as civil society in one view but not in the other.

As a result, independently, these views do not provide me the analytical leverage I need to examine the totality of civil society and its effects on municipal performance in a context where self-supporting organizations are scarce, aid dependence high and the state highly centralized and fused with the dominant party. For example, in the democratization debate, voluntary, self-sustaining non- governmental organizations typify civil society. While I found financial sustainability to be the key characteristic in judging the strength of civil society organizations, most of the organizations I studied were not self-supporting. Similarly, I found that the organizations in market management were extensions of municipal government and, while performing functions on behalf of the municipality, these organizations did represent vendor interests. Under both views, private sector organizations are typically excluded. However, in one case, I found a microenterprise that began as a community organization and eventually received municipal contracts, which extended trash collection services to informal neighborhoods. Lastly, neither view explained how civil society organizations start, grow and change over time and, thus, limited my view of how they interact with the state in ways that affect government performance.

Both views have important implications for municipal performance. Based on the first, I hypothesized that civil society should improve municipal performance because it applies pressure to the municipality to redistribute resources in the form of improved access to services, forcing the state to be

accountable and more inclusive by incorporating citizen input. Based on the second, municipal services should suffer both because there is an absence of effective pressure groups and power is concentrated in a few government elites, potentially resulting in weak accountability. As I discussed, in theory, strong civil society organizations should hold the state accountable. As a result, I expected a strong civil society in Maputo, Beira and Dondo to drive the high municipal performance I observed and a weak civil society in Matola that had no influence, resulting in the low performance I observed in my experiments (see Chapter 3).

Based on this review, I begin my analysis in Chapter 3 with the theoretical supposition that municipalities that have high performance in the test services (trash collection and market management) have civil society organizations that can pressure the municipality to provide better services. I adopt this definition as a starting point because Mozambique today is a new democracy, not a socialist system, with successive elections and a large donor presence vested in making decentralization function. As I

GLVFXVVLQ&KDSWHU,PHDVXUHGWKHVWUHQJWKRIFLYLOVRFLHW\RUJDQL]DWLRQV¶DELOLW\WRSUHVVXUHWKH municipality by examining episodes in which civil society organizations lobby for concessions in municipal policy and that concession is granted by the municipality. In summary, based on my initial review of performance, I would expect Maputo, Beira, and Dondo to have strong civil society organizations in the two test areas and Matola to have weak civil society organizations based on poor municipal performance (see Table 2.1).

Revenue: Earned and Unearned Income (Rents)

Fiscal contract literature examines how different types of revenue affect government performance by triggering varying degrees of accountability.50 Mick Moore posits that a contract exists between the

VWDWHDQGVRFLHWDOJURXSVLQZKLFKWKHVWDWH¶VGHSHQGHQFHRQHDUQHGLQFRPHREOLJDWHVWKHVWDWHWR

50 ,DPERUURZLQJWKHWHUP³ILVFDOFRQWUDFW´IURPMick Moore. 6HH0LFN0RRUH³5HYHQXHV6WDWH )RUPDWLRQDQGWKH4XDOLW\RI*RYHUQDQFHLQ'HYHORSLQJ&RXQWULHV³International Political Science Revi ew, Vol. 25, No. 3, 2004:297-319. 58

SURYLGHVHUYLFHV7KLVGHSHQGHQFH³SODFHVFRQVLGHUDEOHRUJDQL]DWLRQDOEXUHDXFUDWLFDQGSROLWLFDO demanGVRQWKHVWDWHDSSDUDWXV´ 0RRUH $VWKHDUJXPHQWJRHV³GHSHQGHQFHRI

JRYHUQPHQWVRQEURDGWD[DWLRQIRUUHYHQXHLVJRRGIRUWKHTXDOLW\RIJRYHUQDQFH´EHFDXVHLWFUHDWHVD dynamic in which the state is accountable to fee and tax payers because it ear ns its revenue from them

(Moore 2008, 34). Rents are revenue or unearned income that a state does not earn through the provision of services and taxation of individuals; but rather, revenue is generated from sources like natural resources (i.e., diDPRQGVRLO RUDLG:KDWFDQUHVXOWLVD³UHQWLHUVWDWH´ZKLFKKDVOLWWOH

³RUJDQL]DWLRQDORUSROLWLFDOHIIRUWRQWKHSDUWRIWKHVWDWHDSSDUDWXV´ 0RRUH 7RVXPPDUL]H

UHYHQXH¶VHIIHFWRQJRYHUQPHQWSHUIRUPDQFHILUVWVWDWHVWKDWHDUQWKHLUUHvenue from citizens tend to be accountable to those fee and tax payers to provide them services. Second, states that rely on unearned income will not be accountable to fee or tax payers, thereby negatively affecting governmental performance. More to the point, unearned income alone may negatively affect incentives to both generate income and improve government service because the accountability a government would feel to tax payers as the main generators of income is absent.

The first thread of scholarship focuses on how earned revenue creates a state that is responsive to those who pay it. The development of the European state as a result of negotiations between monarchs and land-owners to enable monarchs to finance wars are key findings in this literature (North and

Weingast 1989, Tilly 1992). Bates and Lien explore the fiscal contract of the state by offering that democratic governments develop because monarchs see the benefits of bargaining collectively with citizens over tax concessions (Bates aQG/LHQ 5RVVHYDOXDWHGWKH³WD[DWLRQSURGXFHV

UHSUHVHQWDWLRQ´DUJXPHQWXVLQJWZRPRGHOVDQGFRQFOXGHGWKDWKLJKHUWD[DWLRQUHODWLYHWRJRYHUQPHQW services tends to increase citizen awareness and participation because individuals expect their governments to deliver services when taxes increase (Ross 2004). Compliance with tax collection is more likely when government is perceived to be legitimate and tax payers can see their tax dollars at work (Bräutigam 2008, Fjeldstad 2004). Work done in South Africa has demonstrated that the devolution of charging fees for public services from the central to the local government encourages firm-like behavior among local governments, making them more efficient in the delivery public services 59

DQGUHVSRQVLYHWRWKHLU³FXVWRPHUV´ +RIIPDQ $OOWKHVHVFKRODUVVKDUHLQFRPPRQWKHLU exploration of incentives between the governed and government in maintaining a ³JLYHDQGWDNH´ relationship between the state and tax payers.

The second thread of the literature focuses on how government performs when funded through rents or unearned income. In this study, unearned income refers to both aid and central government transfers to municipalities because both are not generated as a result of services municipalities provide to their citizens. The rentier state literature examines how sources of revenue, which can be generated from natural resources or foreign aid, create conditions in which states do not have to earn revenue by providing services to tax-SD\LQJFLWL]HQVEXWUDWKHUOLYHRIIWKH³UHQWV´WKDWWKHVHVRXUFHVFDQSURYLGH

Scholars hold that this unearned income reduces state incentives to engage with tax payers in what some

VFKRODUVFDOO³UHYHQXHEDUJDLQLQJ´ 0RRUH $VDUHVXOWEHFDXVHRLOVWDWHVIRUH[DPSOHGRQRW generate revenue through a broad tax base but rather through state-controlled resource bases, the state is not incentivized to perform for its citizens.51 An example observed in the Middle East by one scholar showed that ³JRYHUQPHQWVXVHWKHLURLOUHYHQXHVWRUHOLHYH social pressures that might otherwise lead to

GHPDQGVIRUJUHDWHUDFFRXQWDELOLW\´ 5RVV 7KLVFRQFHSWLVDOVRDSSOLFDEOHLQGLVFXVVLRQV about aid as a resource curse in that governments, which enjoy access to large rents, such as those generated by foreign aid, may actually decrease tax collection efforts (Moore 2004, 297-319).

Furthermore, one study of 32 African countries measured how aid affected quality of governance to concludes WKDW³KLJKHUOHYHOVRIDLGDUHDVVRFLDWHGZLWKODUJHUGHFOLQHs in the quality of governance and

LQWD[UHYHQXHVDVDVKDUHRI*'3´ %UlXWLJDPDQG.QDFN While some scholars have

IRXQGWKDWGRQRURYHUVLJKWFDQPDNHDLGOHVVRIVUHQWDQGPRUHRID³VFUXWLQL]HGUHYHQXH´Uesearchers have also found that high levels of aid dependency influence budget policies and processes and weaken

WKH³VFRSHRIUHYHQXHEDUJDLQLQJ´EHWZHHQFLWL]HQDQGVWDWH Collier 2006, Bräutigam 2000, de Renzio

2006, 627-45, de Renzio 2007, Moore 2008, 61). Scholars have found that a connection exists between

51 Work on the effects of oil revenues demonstrates both how oil hinders the development of democracy DQGPD\HYHQEHOLQNHGZLWKWKHLQVWLJDWLRQRIFRQIOLFW6HH0LFKDHO5RVV³'RHV2LO+LQGHU 'HPRFUDF\"´:RUOG3ROLWLFV -361. 60

high aid-receiving states and disincentives for creating sustainable budget structures and seeking citizen input (Bräutigam 2000, Bräutigam and Knack 2004). Recent work in Africa suggests that the absence of aid actually increases accountability. Research in Somaliland demonstrated that the absence of high levels of aid prevalent in Sub-Saharan Africa actually provided Somalilanders leverage over negotiating for more political accountability from the central government (Eubank 2010).

In the context of decentralization, central government transfers can also be considered a type of rent for local governments that produces similar performance disincentives as unearned income at the national level. Recent research in Tanzania and Uganda indicates that districts that receive increased central government transfers have less incentive to generate local taxes (Hoffman 2006; Kjaer n.d.;

Kjaer 2005). In summary, this literature shows that large amounts of aid, little oversight over resources, and low levels of unearned income negatively affect government performance at both the national and subnational levels. Within this context of subnational finance, I observed that municipal budgets for the four cases categorized donor assistance as receipts. So, while these funds were directed, the use of these funds was typically jointly determined between the municipality and donor. As a result, I categorized this municipal-level aid as unearned income.

My research examines two components of the fiscal theory and aid debate. The first is that the fiscal contract research assumes that there are tax lobbies that constitute a commanding share of revenues and that they are able to exercise this revenue influence to bargain with the state. Although scholars have studied how the type of services states provide (social welfare or property rights protection) is affected by how much the state relies on the revenue from a particular group (low-income tax payers versus richer tax payers, respectively), the assumption is that there are dominant groups that can sanction the state by withdrawing fee or tax payments (Timmons 2005, 530-567). Moore offers that governments are generally responsive to major income providers and those that have an exit

option.52 So what happens when there are multiple smaller fee payers, donors, central government transfers and a large informal sector? My research explores whether small tax payer groups can and do sanction municipal government within the context of reliance on central transfers and foreign assistance.

Second, I look for examples of how PXQLFLSDOLWLHVUHDFWWRDLGDVD³UHQW´VSHFLILFDOO\,H[SHFWHGWR find low levels of own-source revenue generation in my cases in an environment of high rents and poor performance. Based on this literature review, I hypothesize that high levels of unearned income in municipalities in the form of aid and central government transfers disincentivize the delivery of municipal services (see Table 1). This hypothesis is the launching point for my discussion in Chapter 5.

Elections: Democratic Performance versus the Performance of Democracy

Many decentralization programs often include a component of political liberalization, making the

GHPRFUDF\OLWHUDWXUHDUHOHYDQWILQDOSRLQWRIUHIHUHQFHIRUWKLVVWXG\&KHHPDDQG5RQGLQHOOLVWDWH³E\ the early 1990s, 95 percent of the countries with democratic political systems had subnational units of

DGPLQLVWUDWLRQRUJRYHUQPHQW´DQGE\WKHUHZHUHPRUHGHPRFUDWLFWKDQQRQ-democratic countries worldwide (Cheema and Rondinelli 2007, 8). Fiscal decentralization scholars see local elections as a

³QHFHVVDU\SUHFRQGLWLRQ´WRVXFFHVVIXOGHFHQWUDOL]DWLRQ %DKO-61). From an accountability

SRLQWRIYLHZGHPRFUDF\LVFRQVLGHUHGWREHWKH³PRVWFHOHEUDWHGLQVWLWXWLRQDOPHDQVIRUFUHDWLQJ

SROLWLFDODFFRXQWDELOLW\´DJRDOHTXDOO\VRXJKWE\SRlitical decentralization (Weingast 2009, 288).

The argument in the democracy literature is that elections affect performance because elected officials compete for (re-)election, compelling these officials to adhere to voter demands for better, more responsive, more accountable government (Dahl 1971; Diamond 1999; Lijphart 1999; Huntington

52 This is a weakness in rational choice institutionalism, that of uneven power distribution between principals and agents, which Terry Moe explores succinctly in his article ³3RZHUDQG3ROLWLFDO ,QVWLWXWLRQV´Perspectives on Politics 3, no. 2 (June 2005): 215-33. 62

1991; Fukuyama 1992) (see Table 2.1).53 More to the point, the introduction of local democracy in the form of decentralized authority for local elections creates space for political competition. As such, local elections introduces another lever which voters can now use to pressure now-electable officials to deliver on campaign promises in order to be re-elected. One study of Indonesian municipalities revealed that the quality and quantity of public services improved because elected officials responded to the threat of sanction elections present (Eckardt 2008, 1-17). In India, the state, seeking electoral support among women and the poor, provided both rice subsidy and micro-credit programs to low- income households (Johnson, Deshingkar, Start 2005, 937-970). In Mexico, municipal government performance responded to high rates of voter participation (Moreno-Jaimes, Carlos 2007, 136-153).

The expression of citizen preference is seen as a core principle of performance-based government. 54

Information about government performance plays an important role in expressing these preferences.

Studies of Mozambican voter attitudes suggest that the more informed voters are about government performance, the more likely they are to vote based on government performance.55 To summarize, democracy in the form of elections affects municipal performance to the degree that it can motivate mayors to fulfill campaign promises of increased service delivery.

53 Democracy scholars have argued that democratic government, defined by elections, facilitates protection of individual liberties, human rights, education and economic development. A core characteristic of democratic government is this choice function²the ability to express preferences through free and fair elections and the right to vote. Robert Dahl, Polyarchy: Participation and Opposition, New Haven: Yale University Press, 1971: 3. 54 ³9RLFHLV«VHHQDVDFHQWUDOO\LPSRUWDQWIDFWRULQWKHPRYHWRZDUGSHUIRUPDQFH-based government (and citizen-RULHQWHGDFFRXQWDELOLW\ ´0DWWKHZ$QGUHZV³9RLFH0HFKDQLVPVDQG/RFDO*RYHUQPHQW )LVFDO2XWFRPHV+RZ'R&LYLF3UHVVXUHDQG3DUWLFLSDWLRQ,QIOXHQFH3XEOLF$FFRXQWDELOLW\"´$QZDU Shah, ed., Public Expenditure Analysis: Public Sector Governance and Accountability Series, Washington, D.C.: The International Bank for Reconstruction, 2005: 220. 55 -RmR3HUHLUD³3DUWLVDQ,GHQWLILFDWLRQLQ$IULFD$Q([SORUDWRU\6WXG\LQ0R]DPELTXH´$WKHVLV submitted to the Faculty of Humanities, University of Cape Town, in fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Political Studies, November 19, 2008. +RZHYHU³XQFULWLFDOFLWL]HQVKLS´UHPDLQVDSHUYDVLYHLVVXHDPRQJ0R]DPELFDQYRWHUV6WXGLHVRI Argentina, South Korea and Japan similarly examine how government performance affects public trust in government. See Soonhee .LP³3XEOLF7UXVWLQ*RYHUQPHQWLQ-DSDQDQG6RXWK.RUHD'RHVWKH Rise of Critical Citizens Matter?, Public Administration Review, September/October 2010: 801-810; 5REHUW0DWWHVDQG&DUORV6KHQJD³µ8QFULWLFDO&LWL]HQVKLS¶LQDµ/RZ,QIRUPDWLRQ¶6RFLHW\ 0R]DPELFDQVLQ&RPSDUDWLYH3HUVSHFWLYH´Afrobarometer Working Paper No. 91, December 2007. 63

Two views challenge the primacy of elections as an effective accountability mechanism that compels performance: the first alternative is that other forms of representation²in this case participatory government²could challenge elections as an effective sanctioning tool to compel performance. Second, elections themselves are open to distortion in decentralization by granting local elites too much power. First, elections alone may be an insufficient mechanism to hold local officials accountable to improving municipal performance resulting in alternative forms of representation²such as participatory governance. Democratic performance is often equated with the quality of elections, not the quality of governance (Diamond, Linz, Lipset 1988). The failure of elections to achieve the touted

LPSURYHPHQWLQJRYHUQPHQWKDVRSHQHGVSDFHIRUDOWHUQDWLYHIRUPVRIUHSUHVHQWDWLRQ7KH³QHZOHIW´ literature examines how participatory governance affects performance when local leadership opens opportunities for participation in the redistribution of resources. Fung and Wright argue that participatory governance enhances accountability, as evidenced by the Porto Allegre case in Brazil

(Fung and Wright 2001, 5-42). Baiocchi and Heller study similar cases in Brazil and India to arrive at similar findings (Baiocchi 2001, 43-71; Heller 2001, 131-163). In Kerala and West Bengal, India, panchayat reforms attempted to create representative democracy and provide fiscal empowerment to individual villages. Low castes were targeted to participate and allocations of funds were made to villages. In these studies, populist politics and an attempt to reorder the power dynamic by including groups of people who were previously excluded from the decision-making process attempted to improve the distribution of public goods in the presence of weak democratic mechanisms.

The second view is that the introduction of elections may actually create an opportunity for local elites to capture of resources, which threatens government performance. Numerous studies suggest that one of the pitfalls of decentralization is that local administration of services without adequate central government oversight and information opens those programs up to elite capture and abuse (Bardhan and

Mookherjee 2000, 135-139; Bardhan and Mookherjee 2006, 101- 2QHVFKRODUQRWHV³LQUHVRXUFH- poor countries, whoever controls the funds can easily manipulate the dynamics of local politics through

SDWURQDJH´ $ZRUWL $V3UDQDE%DUGKDQKDVREVHUYHG³ORFDOJRYHUQPHQWVDUHRIWHQDWWKH

PHUF\RIORFDOSRZHUHOLWHVZKRPD\IUXVWUDWHWKHJRDORIDFKLHYLQJSXEOLF VHUYLFH GHOLYHU\´ %DUGKDQ

2002, 202). Empirical studies in Latin America and Africa demonstrate this dynamic. In Brazil and

Mexico, participatory processes are thought to be an extension of the mayoral spoils systems (Crook and Sverrisson 2008). In his final book, Development and Democracy in Africa, Ake wrote that,

³$IULFDQHOLWHVSODFHGJUHDWREVWDFOHVLQWKHZD\RIGHYHORSPHQWE\WKHLUDQWLSDWK\IRUGHPRFUDF\

(which became) less democratic as its democratic elements, such as the consent of the governed, the accountability of power to the governed and popular participation came under pressure from political elites (emphasis added)´ $NH

While parallel participatory processes may sprout up in a municipality to re-allocate resources or decentralization may institute local elections without adequate central government oversight resulting in elite capture of public resources, the conventional argument continues to focus on how elections incentivize local officials to fulfill campaign promises (see Table 2.1). This argument generated three questions my research addressed. First, elections as a sanction in a de facto one-party state might not prove to be an efficacious sanction because political parties remain underdeveloped and cannot effectively challenge the dominant party. Studying political pressure in a de facto one-party state warrants examination of both intra-party competition for nominations and inter-party competition for elected offices. Second, democracy performance is often linked to the quality of elections rather than government performance. Specifically, the focus in this literature is often on the strength of political parties and transparent elections rather than whether the government elected can deliver citizen services.

Lastly, this literature tends to overlook that how democracies perform by delivering much- needed services to its people requires strong, capable leadership. As a result, leadership and technical aptitude tend to be underemphasized as factors that contribute greatly to municipal performance. That involves viewing the mayor as more than a campaigner earning votes but a C.E.O. instituting change by leading a team that improved municipal performance²whether that is instituting a participatory budget process in Dondo to involve residents in resource-allocation decisions; initiating a fair schedule of scaled trash fees based on electricity usage accepted by Maputo residents; or, regular trash collection in

Beira. Although my research found interesting variation in the degree to which elections incentivized mayors to adhere to their campaign promises, what was most telling was that across high-performing municipalities, mayors were strong C.E.Os. In two instances, mayors had advanced technical degrees whereas in the third, the mayor was a municipal administrator and familiar with municipal administration prior to election. All three built a team of trusted advisors who instituted changes in the collection of own-source revenue and worked closely with donors to get the capacity-building funds to make improve services. These factors were critical in building capacity and communication that led to the municipal performance improvements I observed.

My final finding for two of my cases was unexpected. Holding municipal performance constant, meaning both municipalities were performing trash collection regularly and building markets for vendors, municipal performance had a greater impact on mayoral re-election in the opposition municipality than the dominant party municipality. As discussed in Chapter 1, both Beira and Maputo were high-performing municipalities. As will be discussed in Chapter 6, both mayors ran performance- based campaigns for the 2004 municipal elections and won based on those platforms. They both improved services in their respective cities. When the 2008 campaign season started, both mayors were not re-nominated by their respective parties. However, in Maputo, the FRELIMO municipality, the mayor stepped aside while. In Beira, the RENAMO municipality, the mayor ran as an independent on a results-based platform, won and started a new national party that has since gained seats in the

Parliament in the 2009 national elections.

Based on the work of democracy scholars I previously discussed, my starting point in Chapter 6 is based on the premise that elections improve government performance because mayors compete for votes. Therefore, I hypothesize that competitive elections incentivize mayors to improve municipal performance to constituents because mayors seek to be re-elected. As discussed in Chapter 1, given that

Mozambique is a de facto one-party state, I defined political competition as the degree to which there was party and/or mayoral turnover and the number of parties that competed in those elections. I argue that the more competitive an election²meaning the more competitors and the greater political space as

measured by both mayoral and party turnover²the more re-election means to a mayor and the harder he will work to deliver on his/her campaign promises. Based on performance and electoral results since

1998, I would expect Beira to have highly competitive elections, Maputo to have a medium level of competition, and Dondo and Matola to have a low level of competition (see Table 2.2).

Conclusion

As this chapter demonstrated, the literature on decentralization identified that civil society, revenue, and elections were three factors that affect performance (see Table 2.1). The literature on civil society indicated that a strong civil society holds government accountable through increasing citizen participation and voice to pressure municipal government to improve services. Studying a post-socialist, low-income new democracy required a rethinking of the concepts of civil society to adapt to the

Mozambican context. In a context where institutions are new and fragile and resources are limited, civil society organizations became sustainable fixtures in municipalities as service providers, lobbies, or, eventually, microenterprises, as was the case in Maputo. In Beira, they contracted until resources become available. In Beira, Dondo and Maputo, civil society organizations are extensions of the state, not as representatives of the elites but facilitators between the government and the people, a role not uncommon in former socialist and communist countries. From this point, I hypothesize that strong civil society organizations will drive improvements in municipal service in Maputo, Beira and Dondo and expect that civil society will have no measurable influence in Matola (see Table 2.2). In Chapter 4, I examine my evidence to determine if this theory is valid.

Municipal performance also relies on available resources. The fiscal contract literature provided me a framework for analyzing revenue incentives and performance. Earned income²or own-source revenue²should encourage municipal governments to provide services because residents are

FRQVXPHUV³SD\LQJ´IRUWKHVHVHUYLFHV8QHDUQHGLQFRPH² like aid and, in this case, central government transfers²should discourage municipal governments from improving services because the

consumer incentive is missing. In the context of Mozambique, where it is much less likely that there is fee or tax lobby that holds a commanding share of revenue, there was an under-emphasis of how multiple, small fee payers bargain with the state. Additionally, the literature underemphasizes how aid as unearned income can and does have a positive role. So, I begin Chapter 5 with the premise that high performing municipalities rely on earned income while low performing municipalities to rely on unearned income (see Table 2.2).

Lastly, I examine how elections affect performance in Chapter 6. Democracy scholars focus on how elections improve government performance because, like the dynamic of earned income discussed

LQWKHSUHYLRXVVHFWLRQHOHFWLRQVSURYLGHYRWHUVD³FRQVXPHU´RSWLRQ²they can reward their elected official with a vote because of his or her ability to improve municipal service provision. I have found that this view does not provide the analytical leverage to examine political pressure exerted by elections in a de facto one party state. Furthermore, the recurrent theme in this discussion is that democracy performance is measured by the quality of eOHFWLRQVDQGWKHPD\RU¶VDELOLW\WREHUH-elected instead of the quality of government performance. As a result, too little attention is paid to the qualities elected officials must have as managers to run a bureaucracy. I undertook this research to examine not only how mayors campaigned but also what they did in between campaigns and how these actions contributed to overall municipal performance. I begin Chapter 6 expecting that competitive elections result in mayors that can elevate the quality of municipal performance because they are incentivized to do so (see Table 2.2).

My final chapter reframes and extends the implications of my findings in the previous chapters beyond decentralization to the fields of development studies, public administration and political science more broadly. It demonstrates that, even in the context of weak institutions and serious constraints, local governments are able to build revenue capacity and sustainable political independence in one-party states. The conclusion also recommends areas for further study.

Table 2.1. Hypothesis and Findings

Hypothesis: Decentralization creates incentives that motivate municipal government officials to act and opportunities for residents to pressure them.

Resource pressure over local officials increases as new authorities put resources under the control of local decision-makers.

Revenue:

Expected Finding: Municipalities with majority shares of earned income will have high performance in test areas while municipalities with majority shares of unearned income will have low performance. In Mozambique, high levels of unearned income in the form of aid would lead me to believe that all four municipal cases would have low performance.

Finding: Own-source revenue (vendor and trash fees) grew in the presence of majority shares of unearned income. Government performance improved in the absence of a tax lobby. Own-source revenue growth alone does not improve performance. Unearned income improved municipal performance in cities with strong mayoral leadership.

Political pressure over local officials increases through new opportunities for expression of citizen preferences through voting and civil society participation.

Civil Society:

Expected Finding: Strong civil society can pressure municipal government to improve municipal services.

Finding: Strong civil society is not a prerequisite to strong municipal performance. Civil society organizations are fluid and adaptive to resource availability. They affect municipal performance based the type of donor assistance it receives and its relationship with the municipality.

Elections:

Expected Finding: Competitive elections incentivize mayors to improve municipal services to constituents.

Finding: Dominant party politics affected the ability for voters to exert political pressure through votes. Technocratic mayors improved performance. Mayoral performance and constituent votes mattered most for re-election in the opposition municipality.

Table 2.1. Theoretical Launching Point, by Municipality, Service and Variables

CIVIL REVENUE ELECTIONS MUNICIPAL SOCIETY PERFOMANCE

MAPUTO

Markets and Trash Strong Earned Medium High Collection Income Competition

BEIRA

Markets and Trash Strong Earned High High Collection Income Competition

DONDO

Markets and Trash Strong Earned Low Competition High Collection Income

MATOLA

Markets and Trash Weak Unearned Low Competition Low Collection Income

III. USING EXPERIMENTAL APPROACHES: DEMONSTRATING VARIATION ON THE DEPENDENT VARIABLE

Nobody really knows who owns what where, who is accountable for the performance of obligations, who is responsible for losses and fraud, or what mechanisms are available to enforce payment for services and goods delivered. Hernando de Soto (de Soto 2000, 32)

Introduction

Chapters 1 and 2 lay the theoretical and empirical groundwork of this inquiry. A critical component in tracing accountability for municipal provision of goods and services is observing the actual provision of said goods and services. The purpose of this chapter is to observe the two test services²trash collection and provision of vendor services²to confirm empirically there is variation in my dependent variable²municipal performance. I do this by adapting two experiments from public administration, development, and democracy literatures (Hatry 2007; DeSoto 2000; Putnam 1993). The two experiments are: application for ambulatory business licenses in three of the four cities and trash collection monitoring over a specific period of time both in an informal and a formal neighborhood in all four cities.56

Decentralization vested municipalities with nine key functional responsibilities ,which include granting ambulatory business licenses and trash collection. 57 I chose these two services because they are wholly municipal responsibilities under decentralization, municipalities have the authority to keep

100 percent of the revenue generated from these activities, and they are easily observable. Furthermore, both these services vary in type as one is a transactional service between one payer and the municipality

(ambulatory vendor licenses) and the other is a public good (trash collection). For ambulatory licenses,

56 Dondo was excluded because maintaining anonymity to conduct the exercises was not possible.

57 Municipal responsiblities under decentralization include installation and maintenance of rural and urban infrastructures, such as markets and fairs and cementaries, basic sanitation, distribution of electricity, transportation and communication, education, parks and recreation, health, care of vulnerable populations, and environmental management. Teodoro Andrade Waty, Autarquias Locais: Legislação Fundamental, ML Graphics: Maputo, 2000, 111-112. For more about decentralization, please see Chapter 1.

the quality and benefit of the service rendered accrues to one individual and are in essence private benefits. Additionally, this transaction is open to corruption because of the direct consumer relationship between the vendor and municipal official processing the license. Trash collection is different because it is a group benefit, and, as a result, a public good. Unlike business licenses, trash collection is a more complex service because it requires extensive infrastructure investment and the coordination of behavior among residents to maintain cleanliness, unlike the singular ambulatory license applicant. Lastly, as a group service, trash collection is less likely to be open to corruption (at least between the payer and the municipality) than business licenses. The main challenge municipalities face in providing these, as well as other, services is both the resources and the varying nature of these responsibilities.

I begin this chapter by examining the evidence on the independent variables to posit the expected relationship between these factors and municipal performance. Then, I review the results of my ambulatory vendor experiment in Maputo, Beira and Matola, which demonstrate that there is indeed variation in the quality of services. For trash collection, Maputo, Beira and Dondo perform well, with

Beira and Dondo being best able to keep up with the flow of trash production and Maputo struggling, especially in the formal neighborhood, to keep up with the flow of trash. Matola performed less adequately in comparison, with erratic and incomplete pick-up patterns. For ambulatory business licenses, Maputo outperformed Beira and Matola with the most professional, efficient system in place

IRUSURFHVVLQJDSSOLFDWLRQVLQFRPSDULVRQWR%HLUD¶VVRPHZKDWRSDTXHEXWDGHTXDWHSURFHVV0DWROD¶V process was the least efficient and most expensive. These results confirm that there is variation in the dependent variable with high municipal service in three municipalities and low service in the fourth.

Initial Evidence: Summary of Independent Variables by Case

Maputo

7KHFDSLWDO¶VQDPH0DSXWRFRPHVIURPDWUDGLWLRQDO5RQJDGULQNQDPHG8SXWVRPDGHIURP fermented prunes, which is arguably more than 1,000 years old (Moçambique 2005). Maputo is

situated on Maputo Bay to its South and East and the Infulene River to the West, in Southern

Mozambique, approximately 100 kilometers from the South African border. The highway connecting

Maputo to South Africa, along with the railway line and Port of Maputo are the most important means of transportation for export markets in tKHFRXQWU\7RGD\¶VSRSXODWLRQLVURXJKO\RYHUPLOOLRQ inhabitants, divided in seven urban districts covering 64 neighborhoods with 13 percent in the urban districts living in the cement part of the city and 81 percent living in the suburban areas (CMM 2008).

'LVWULFWLVFDOOHGWKH³FHPHQWFLW\´EHFDXVHLWLVWKHROGSDYHGSDUWRIWKHFLW\ZLWKWDOOEXLOGLQJV broad, tree-lined boulevards, parks and the seat of all central and municipal government buildings,

LQFOXGLQJWKH3UHVLGHQWRIWKH5HSXEOLF¶VKome, donors and embassies.

$VWKHFDSLWDO,ZRXOGH[SHFWWKDW0DSXWR¶VFLYLOVRFLHW\WREHDFWLYHGXHWRWKHSUHVHQFHRI funders such as ministries and foreign donors, commercial activity to be high so municipal revenue would be as well, and the FRELIMO party would be strong as the seat of national government. Maputo is the national seat of many donors and non-profits, such as the Fundação de Desenvolvimento

Moçambicano. Of the 301 organizations in a 2008 publication of development organizations (includes local NGOs, government organizations, and donors), 90 percent are located in Maputo with the remaining 10 percent located mainly in Beira, Nampula, Pemba, and Chimoio (Directory of

Development Organizations 2008). In terms of resources, Maputo is doing well. Municipal revenues and aid in Maputo hover around the average for all 33 municipalities as measured on a per capita basis.58 I would also expect aid to play a larger role than other sources of revenue because of volume of aid. On politics, Maputo voters have consistently elected FRELIMO mayors and FRELIMO-majority municipal assemblies; however, it has had consistent turnover on all three elections in mayor. As a

UHVXOW,ZRXOGH[SHFWWKDW0DSXWR¶VPXQLFLSDOSHUIRUPDQFHLQWKHDUHDVRIWUDVKFROOection and vendor services to be comparatively high, indicating some intraparty competition.

58 Non-aid revenues (own-source and central government transfers) average from 2000 to 2008 is $6.40 USD and aid per capita averages from 2004 to 2007 are $18.53 USD.

Beira

7KHZRUG³EHLUD´LQ3RUWXJXHVHPHDQV³WKUHVKROG´RU³ERUGHU´,QPDQ\ZD\VLWLVZLWKWKH

FRXQWU\¶VVHFRQGODUJHVWSRUWDQGRQHRIWKHWKUHHPDMRUUDLOZay arteries connecting Beira to

Zimbabwe.59 The city of Beira, capital of Sofala province in central Mozambique, was named after the

Ponce of Beira, D. Luis Filipe, the eldest son of Portuguese King D. Carlos I in July 27, 1887 (Folha

Informativa 2002; Cem anos 2007). Situated about 247 kilometers from the Zimbabwean border, Beira, the capital of Sofala province in central Mozambique, is situated on slip of land that juts into the Indian

Ocean to the South and River Pungué to the West and the district of Dondo to the North. Under sea level, Beira is often prone to flooding during rainy season, with a hot, tropical climate that incubates lush palm fronds, rice fields, and coconut, almond, and papaya trees. The city spreads its 26 neighborhoods, the majority of ZKLFKDUHORFDWHGRXWVLGHWKH³FHPHQWFLW\´RYHUVTXDUHNLORPHWHUV

(Folha Informativo 2002). Beira is the third most populous city, with 436,240 residents, and boasts a port, railway and maritime shipping industry of regional and international significance (Census 2007). It is also the location of one of the countries three major hospitals.60 %HLUDKDVWKHFRXQWU\¶VVHFRQG largest industrial park, behind Matola (Folha Informativo 2002). In comparison to Maputo, Beira has a sleepy quality. It is common to walk along the paved roads, observing the Ponta Gêa Cathedral while

KHDULQJWKH0XVOLP³FDOOWRSUD\HU´

$V0R]DPELTXH¶VFHQWHURIWKHRSSRVLWLRQ,ZRXOGH[SHFW%HLUDWREHDORZ-performing municipality primarily because it is a RENAMO stronghold and, as such, not in control of government resources, leading to an inability to invest in infrastructure and services. I would assume that the

FRELIMO-dominant central government, which not only controls transfers but the majority of foreign

59 Known as the CFM-Centre (Caminhos de Ferro de Moçambique), CFM South is Maputo which conntects Maputo to Swaziland, South Africa and Zimbabwe and CMF-North connects Nampula to 0DODZL81'3´0R]DPELTXH1DWLRQDO+XPDQ'HYHORSPHQW5HSRUW+XPDn Development to 2010-5HDFKLQJIRUWKH0LOOHQQLXP'HYHORSPHQW*RDOV´0DSXWR0R]DPELTXH

60 Maputo and Nampula are the two other cities.

assistance, would penalize Beira by both withholding transfers and delaying aid disbursements which could have two effects. First, without timely funding, the municipality cannot provide services or collect taxes. Second, civil society groups are weakened without financial support.

However, these weaknesses are balanced against an active political atmosphere and a strong mayor that resulted in a high-performing municipality. Mayor Simango improved trash collection greatly in Beira and Beira voters were able to reward him for it. They demonstrated more personal, than party, loyalty in the way they voted, resulting in unique variation which catapulted Daviz Simango not only to re-election as an independent in 2008 after being thrown out of the RENAMO party but to the creation of his own national party. Despite being an opposition municipality, Beira received a higher per capita average of revenue and twice the amount of aid per capita than Maputo.61 Leadership and the availability of resources might account for improvements in municipal services. So, I would expect to find functioning trash collection and ambulatory vendor license processes in place.

Dondo

Dondo is the capital of the Dondo district, 30 kilometers southwest of Beira and about 300 kilometers east of the Malawian border. An interior, agricultural hamlet, Dondo is surrounded by

Mafambisse to the North, Beira to the South , Chinamacomo to the East and Buzi to the west. The word

³GRQGR´LVEHOLHYHGWREHDFRUUXSWLRQRIWKHQDWLYH³GKRQGR´ZKLFKPHDQVWKLFNMXQJOH 5RTXHDQG

Tengler 2000). The first modern settlement was founded in 1891 as a result of the British South Africa

&RPSDQ\¶VHVWDEOLVKment of a railway station in Dondo. Dondo was briefly annexed to the City of

Beira in 1933 and reverted back to its own administration in 1955. Today, Dondo has 61,345 residents, with the majority of population resides in ten neighborhoods in the urban center.

61 )URPWR%HLUD¶VDLGSHUFDSLWDLVVRKLJKEHFDXVHRIDPDMRUPLOOLRQ:RUOG%DQN water/sewer infrastructure project which constitutes 41 percent of the total aid for this four-year period.

$VD)5(/,02PXQLFLSDOLW\LQDSURYLQFHGRPLQDWHGE\5(1$02,ZRXOGH[SHFW'RQGR¶V municipal performance to be high because of political considerations, available resources, and mayoral leadership (see Table 3.1). Dondo is the largest FRELIMO municipality in Sofala province, 30 minutes away from RENAMO-dominant Beira. As a result of these political considerations, I would expect

'RQGR¶VSHUIRUPDQFHWREHKLJKIRUWKUHHUHDVRQV)LUVWWKH)5(/,02-dominant central government would be politically embarrassed if this municipality did not provide good services and neighboring

RENAMO-PXQLFLSDOLW\%HLUDGLG$VDUHVXOW,ZRXOGH[SHFW'RQGR¶VUHYHQXHWRGHULYHSULPDULO\ from the central government and for the municipality to have a low share of own-source revenue to overall revenue as a result because the incentive to collect own-source revenue would be low. Average per capita, excluding aid, showed that Dondo receives the highest average of the four cases

(USD$9.54/resident). Second, Dondo is a national model of participatory budgeting so I would expect

'RQGR¶VSHUIRUPDQFHWREHKLJKEHFDXVHLQDGGLWLRQWRWUDQVIHUV'RQGR¶VDLGUHFHLSWZDVKLJK)URP

2004 to 2007, Dondo received one and a half times the average aid per capita (USD$28.35/resident). I would also expect municipal performance to be high because Dondo has had the same mayor since

1998, providing consistent leadership for ten years. Lastly, because Dondo is a national model of participatory government with high levels of aid, I would expect civil society to be developed and positively influence municipal performance.

Matola

Modern-day Matola, situated 12 kilometers from Maputo on route west to South Africa, was

formerly part of Maputo. It is the capital of Maputo province and the industrial capital of

0R]DPELTXHZLWKSHUFHQWRIWKHFRXQWU\¶VLQGXVWULDOSDUNVFRQFHQWUDWHGZLWKLQLWVFLW\OLPLWV

)ROKD,QIRUPDWLYD )RXQGHGDVDYLOODJHLQ0DWROD¶VSRSXODWLRQRILVGLYLGHG

among three urban districts that contain 41 neighborhoods and cover over 375 square kilometers

(Ibid). As a FRELIMO stronghold, I would expect Matola to have low performance for the

following three reasons. First, unlike Dondo, which is seen as a provincial political counterbalance

to Beira, and Maputo, a capital with a heavy international presence, there is no incentive for the

FRELIMO-GRPLQDWHGFHQWUDOJRYHUQPHQWWRLQYHVWLQ0DWROD¶VPXQLFLSDOSHUIRUPDQFH VHH7DEOH

3. 0DWROD¶VUHYHQXHSHUFDSLWDDYHUDJHVIURPWRH[FOXGLQJDid, was the lowest

revenue per capita average of the four cases (USD$2.42/ resident). Second, I would also expect

that Matola was often overlooked by donors because they tended to focus more on neighboring

Maputo than Matola. The average aid per capita, the lowest of the four cases at USD$3.18/resident,

was consistent with this observation. The lack of resources from both central government transfers

DQGDLGFRXOGDOVRLQGLFDWHWKDWFLYLOVRFLHW\ZDVZHDNGXHDODFNRIUHVRXUFHV/DVWO\0DWROD¶V

leadership was unexpectedly disrupted. Despite the fact that Matola voters elected the same

FRELIMO mayor since decentralization, his second term was cut short by his unexpected death,

when he was replaced by the Municipal Assembly President. Another FRELIMO candidate was

elected in 2008.

Ambulatory Vendor Experiment

In order to assess municipal performance of a basic service for vendors²provision of an ambulatory vendor license²my research assistants applied for ambulatory vendor licenses in Maputo,

Beira, and Matola. The purpose of applying was to determine how long it would take, how much it would cost, how effective city officials were in providing timely and accurate information, and how many steps the applicant would need to take into order to receive the license (see Table 3.2). It is also the easiest way for a person to enter into the vendor trade, as getting a desirable space at the major markets was challenging. Similarly, ambulatory vendors, like many truly informal vendors, can literally

³FKDVH´the customer and it requires little start-up investment. Lastly, licensing these activities is a municipal responsibility that requires little effort compared to managing the markets because it is a matter of simply processing applications and tracking vendors, unlike market management, which involves paying staff to manage infrastructure and balancing relations among vendors. So I determined

that this would be a good area to conduct an application experiment as the application process for the vendor should be relatively straight-IRUZDUGDQGWKHFLW\¶VUHVSRQVHHTXDOO\VR0DSXWR¶VSURFHVVZDV the most efficient, taking only 22 days to receive the license in 6 steps. Beira took 38 days and 8 steps with Matola requiring 55 days and 8 steps to receive an ambulatory license. Table 3.2 demonstrates the steps, days, and stops the applicant made to receive the license.

Maputo

Unlike Matola and Beira, ambulatory vendors are ubiquitous in Maputo. Its population and density make selling in the streets a relatively easy, quick way to earn money. Of the three cities,

0DSXWR¶VSURFHVVZDVWKHPRVWHIILFLHQWWDNLQJGD\VWRUHFHLYe the license from the date of application and requiring the applicant to go to only three locations, two of them co-located, and undertake only six steps. Technically, the vendor can sell his or her goods the day the fee payment is made, using the receipt as proof. Maputo was the only city to streamline the process in this way. Receipt of the license itself took 22 days.

Overall, the process was highly professionalized. The applicant was able to get the total price for the license, all the information that was needed to apply and an estimated length of time until the license would be completed, which was 15 days, not far from the actual time of receipt, at the first visit.

Furthermore, the CMM issued application-tracking numbers, which the applicant could use to monitor progress. When the applicant called to follow-up, the attendant picked up and was able to track the application and inform the applicant. The applicant did not have to unnecessarily return to any of the three sites, all of which were organizationally within the Directorate of Markets and Fairs, and two of the sites where physically co-located in the same building. Lastly, the applicant received both a license and an identification card. Maputo¶s process was the most transparent and efficient of the three cases.

+RZHYHU0DSXWR¶VIHHVFRVWDQLQGLYLGXDODWRWDORI07QRUWKHHTXLYDOHQWRIDSSUR[LPDWHO\

USD$45, a large amount considering the average per capita income.62 The cost (it was the second most expensive of the three) might still be prohibitive enough to explain why so many ambulatory vendors remain dumba nengue or informal vendors.

Beira

2IWKHWKUHHFLWLHV%HLUD¶VSURFHVVZDVWKHVHFRQGPRVWHIILFLHQWLQFRPSDULVRQWR0DSXWR taking 38 days and 8 steps, involving three city offices. Although Beira was the second most efficient in terms of the measures, it was the most circuitous process of the three and involved follow-up in different locations, which adds cost and effort to the applicant, thereby discouraging him to apply. The most discouraging part of the process was that the applicant could not get all the information he needed to apply upfront. This was particularly problematic when the applicant needed accurate information about costs. The applicant had to unknowingly pay various fees along the way, which both potentially opens the process up to corruption, bribery and revenue escaping the public treasury into individual

KDQGVDQGLQKLELWVDQDSSOLFDQW¶VDELOLW\WRVDYHIRUWKHOLFHQVH

Overall, the process was opaque. When the applicant first spoke with the Directorate of

Industry, Commerce and Tourism, he requested to sell clothes, which the regulations do not permit.

The regulations were so outdated that licensed ambulatory vendors can only sell dried fish and shrimp.

Although the city of Beira is the largest of the four cities under study in terms of surface area, making it a more dispersed city with less foot traffic, the CMB is missing out an opportunity for revenue by appearing to enforce an outdated law. I observed ambulatory vendors selling clothes and cleaning supplies who were clearly not working with a license. Although the applicant was told in general terms what he would need, he was not told how long the process would take and the prices for fulfilling the requirements of the license kept changing throughout the process. The process was circuitous. The

62 This charge includes the personal municipal tax or IPA (Imposto Pessoal Autárquico). This cost reflects not just the charge of the CMM but also the cost to the individual, which includes paying for items reasonably required for the process such as pictures and chapa fare.

applicant had to report to the Directorate of Health in order to get examined and get a health card to enable him to sell food products. He was told that the charge would be 23 MTn (less than USD$1.00) and it would take a month or he could pay 125 MTn, he would not have to be examined, and the health card could be processed the same day. When he returned with the 125 MTn and paperwork, he was able to get the health card that same day without an exam. From there, he had to report to the Secretary of his bairro to get proof of his residence, which he attempted to do but the Secretary was unavailable. He returned the next day, paid 45 MTn and was given a declaration of residency. Due to the Christmas holidays, the applicant was told to wait a week to turn in the documentation. When he returned he was instructed to pay an additional 388 MTn processing fee, which he was not prepared to pay. He returned two days later and was told he had to pay a higher rate (491.40 MTn) than he was prepared to pay and he received the license the same day. The CMB did not provide a tracking number to the applicant and throughout the process, the applicant was forced to pay ever more money unexpectedly in order to get his paperwork moving. In total, the applicant paid USD$25, which, although costly, is the least expensive of the three cases. The process discourages applicants because of lack of accurate, timely information, availability of city staff, and multiple paperwork requirements opening the process up to bribes. However, the incentives for the CMB to undertake a major review of its policies were weak because the population density per square kilometer is low in comparison to other municipalities in the study. This could have disincentivized vendors from selling in the streets outside of the immediate

PDUNHWDUHDVEHFDXVHWKHUHDUHQRWDVPDQ\FRQFHQWUDWHGFRQVXPHUVYHQGRUVFDQ³FKDVH´

Matola

0DWROD¶VSURFHVVZDVWKHOHDVWHIILFLHQWDQGPRVWH[SHQVLYHRIWKHWKUHHFLWLHV0DWROD¶V application required cost USD$58 and took 55 days and 8 steps. The one benefit of this process in comparison to the others was that the applicant only had to apply in one office although the office was in the process of moving and so he had to report to two different locations that were not co-located.

/LNH0DSXWR0DWROD¶VSURFHVVZDVIDLUO\VWUDLJKWIRUZDUGDQGWKHUHZHUHQRVXUSULVHIHHVDWWDFKHGDW

various stages of the application. However, unlike Maputo, the process took a long time and the level of customer service in Maputo was not present in Matola, making tracking the process extremely difficult.

Unlike Beira, the applicant was able to find out beforehand what he needed to apply (two photos, a request directed to the Mayor )and the cost (1,550 MTn (USD$58.49)). He was able to turn in these items at the Office of the Town Councilor for Markets and Fairs. Four days later, he inquired what the status was on the application and a staff member told him it was in process. Two days later, which coincided with the beginning of the month of December, the applicant checked in again and was told the ambulatory licenses for November had not yet been issued. Eleven days after the application was made, the applicant called to see if the license was ready and he was told that the town councilor had not signed the paperwork yet. Over the next 23 days, the applicant called once and returned twice to inquire regarding the status of the application. He was told either it was not ready or, in one instance, was directed to an office where no one was present because everyone left to attend a speech the Mayor was giving. Finally, after 55 days the applicant returned in person and received the license. Although he was able to get information over the phone about the status of the license, which was not an option in

Beira, the applicant was directed to empty offices and waited almost two months after the initial date of his application, paying USD$58 for a process that took twice as long as Maputo and was the most expensive.

Summary of Ambulatory Vendor License Experiment

The three cases demonstrated variation on the level of municipal performance of processing ambulatory vendor licenses. First, these processes are governed by each municipality that has its own process set up through similar but not identical agencies. Second, the requirements were different. In

Beira, municipal regulations were so outdated that prospective vendors could only (legally) sell dried shrimp whereas in Maputo and Matola, there were very little restrictions on what could be sold by

ambulatory vendors. Although Maputo was the only municipality to recently update its regulations,

0DWROD¶VUHJXODWLRQVGLGQRWGDWHEDFNWRFRORQLDOWLPHVDVGLG%HLUD¶VZKLFKGLGDOORZYHQGRUVPRUH flexibility in what they could sell. In Beira, applicants were required to undergo a medical clearance whereas this was not a requirement in Maputo or Matola. In two cases (Maputo and Beira), proof of residency was required but in the third case (Matola), residency was not a prerequisite. In Maputo and

Matola, the municipalities used the application as an entry point to collect the Municipal Personal Tax that Beira did not do. In all three instances, the town councilor was required to sign the license. Lastly, enforcement of licensing in all three cities is weak to non-existent so there is no sanction to hold ambulatory vendors accountable to having a license. With the cost and length of process, I too might decide to forgo a license.

Although cost and process can be prohibitive factors in applying for a license, all staff I interviewed during and after this experiment concurred that the biggest obstacle to providing better service in this area was cultural²from both municipal staff and vendor perspectives. The culture of customer service among public employees, who tended to have a low limited level of education, few resources, and a tradition of centralizing information as a result of both colonial and socialist influences, was still developing. I observed that Maputo was an exception in part because of the overwhelming presence of international donors. )RUH[DPSOH0DSXWR¶VPDUNHWUHJXODWLRQVZHUHWKHPRVWUHFHQWO\

XSGDWHGRIWKHIRXUFDVHV0DUNHWLQWHUYLHZVLQ0DSXWR¶VODUJHVWIRUPDODQGLQIRUPDOPDUNHWVUHYHDOHG that the fee levels, which were updated in the regulations, were consistent. While one market chief indicated that the CMM had not effectively consulted with the market chiefs, they had received copies of the regulations before they had been final and had an opportunity to comment. I attribute the consistency in reporting fees in Maputo market to the dialog required in updating these regulations, a process funded through the PROMAPUTO project.

Matola and Beira generally have processes that are non-transparent and a propensity for funds

WR³JRZDONLQJ´It is cheaper and easier to simply decide to sell on the street versus trying to get a license when the vendor has something to sell. There is a culture of informality in which consumers are

used to purchasing and vendors accustomed to selling on the street. I attribute this to a combination of lack of formal employment, increased rural migration during the war, a shift from large agricultural producers to small subsistence farming, an influx of Zimbabweans and South Africans which increase the numbers of individuals selling on the streets and a tendency for consumers (anywhere) to look for convenient ways to shop.

Trash Observation Experiment

The second municipal service I studied for variation on my dependent variable was trash collection. In order to measure performance of this service, I conducted a trash observation experiment, an exercise undertaken to measure performance in public administration (see Table 3.1). The purpose was to observe the volume of trash in and around these containers. If trash containers had the same amount of trash every day, I deduced that the municipality was able to keep up with the flow of garbage that meant that performance was fair. If the container and surrounding area were empty and clean, I deduced that the municipality was able to perform an exceptional job by adapting to garbage flows throughout the neighborhood. If the container and surrounding area showed signs of garbage accumulation, then I deduced that the municipality was unable to collect trash at the level needed by the

QHLJKERUKRRG¶VUHVLGHQWV7KLVH[SHULPHQWLQYROYHGVHOHFWLQJRQHIRUPDODQGRQHDGMDFHQWLQIRUPDO neighborhood for observation. I proceeded to select within these neighborhoods a route that contained either containers or concentrations of trash that I could observe and then selected those as my units of observation. I then determined the frequency of service based on interviews with municipal staff and associated materials. In some cases, collection was daily and in others, weekly. I, or my research assistant, photographed and annotated the garbage accumulation at each observation site at the same time every day (or week, depending on the period of service).

Maputo was the only city in which collection was daily and specific to the neighborhoods I observed, meaning that the trash collection occurred 24 hours a day. Dondo and Beira had city-wide

trash collection service ; however, it could be once or twice a week (or as often as the Secretary of the bairro calls the municipality to get service) in neighborhoods under observation. Matola service was weekly in the two neighborhoods under observation. Although Maputo had the most frequent service and was the only municipality with private contractors performing part of the service, the CMM could not keep up with the flow of trash, especially in the formal neighborhood. Beira and Dondo had fairly

UHJXODUWUDVKVHUYLFHWKDWNHSWXSZLWKWKHIORZRIWUDVKZKHUHDV0DWROD¶VZDVWKHOHDVWFRQVLVWHQW

Maputo

District 1

Overall, although service was apparent based on the removal of trash, the levels of trash remained the same. In limited instances, as in the case with the EGF Neoquímica service, trash accumulated and remained uncollected for days.63 At site 1, only two of the three containers were available as the third was damaged beyond repair. The remaining bins were about 1/4 to 1/2 full. By observation day 5, both containers were full; however, day 5 fell on a Sunday, which might explain the increased volume of trash. By day 7, the bins were between 1/3 to 1/2 full, with cardboard boxes separated, the work of the informal collectors who live at this site. Site 2 consisted of one container that consistently had garbage strewn around it. I saw a steady accumulation of trash throughout the week; however, by Sunday, the container had been partially emptied. Site 3, like the previous two sites, also had individuals sorting through the trash looking for recyclables, like cardboard, four of whom appeared to live at the site. Again, trash steadily accumulated, with a reduction of trash on observation day 7, which fell on a Tuesday. Site 4 consisted of three bins. Again, trash gradually accumulated but by day

5, which fell on a Sunday, the bins were completely emptied.

63 In District 1, I examined 6 trash containers: all on Vladimir Lenin Avenue, which I designated based on the cross streets, starting from Frei Tomas Road (site 1) heading southwest towards Maputo Bay and ending at Vladimir Lenin and July 26 Avenue (site 6) Unlike the area observed in District 3, parts of this area are serviced by EGF Neoquímica, a CMM contractor.

Sites 5 and 6 were the only sites under observation in which a CMM contractor, EGF

Neoquímica, did the collection. The CMM contracted this service out in areas of high trash production and these containers are indeed placed in primarily commercial areas of high foot traffic and bus stops.

Although the sites were serviced, both sites were overflowing with garbage that sat for at least two days.

By observation day 2, receptacles at sites 5 and 6 were already overflowing. By observation day 4, trash was collected with the area around the containers relatively free from litter. It remained this way at site

5 through the remainder of the observation period. Site 6 became inundated by day 6 and by day 7, trash was overflowing several feet around the containers.

District 3

6XUSULVLQJO\WUDVKFROOHFWLRQVHUYLFHLQ0DIDODOD¶VPDLQURDGZDVEHWWHUFRPSDUDWLYHO\WKDQ service in the area in District 1 under observation because, I speculate, the pedestrian traffic and commercial activity is not as great. District 1 residents have more disposable income and therefore purchase more prepared goods, creating more trash, and the volume of trash production is simply higher than the capacity to provide the service.64 The first container (site 1) was full of garbage on the first observation date but consistently level between 1/2 to 2/3rds full throughout the week. With the exception of observation day 4, where a garbage pile had accumulated to the right side of the container, the area remained relatively clean. This area is a place of heavy activity, a chapa stop and the first garbage deposit point as one turns off another major road, Vladimir Lenin Avenue, which might explain the consistent presence of trash. The container at site 2 was full of trash that had begun to accumulate to its side. However, by day 2, I witnessed a cleaning crew of 5 individuals, clad in protective coats, sweeping and gathering the trash around the container even though the container itself was still fairly full. By day 4, the container was between 1/2 to 3/4ths full, indicating that the truck had passed and for

64 In District 3, I examined 5 trash containers: three along Chissano Avenue (sites 1,2, and 3), one at the October 21 Plaza (site 4) which leads to Xipamanine Market, and one at Marien Ngoubi Avenue (site 5) which connects Mafalala to District 1. The experiment began on a Sunday and so, I would expect to find copious amounts of garbage that had accumulated throughout the weekend.

the remaining days, the container remained at that level, with the surrounding area swept clean and free from debris. At site 3, which was the largest container of the five observed, the area around the container remained consistently clean with the contents occupying about 2/3 to 3/4ths of the container.

Site 4 was similarly clean, with a consistent amount of trash throughout the week in the container that stayed at the same levels until Saturday, in which the container filled up and trash was strewn on the surrounding ground. I did see a trash worker, clad with a uniform and gloves, on observation day 3 sweeping the trash surrounding the container. On most days, the area surrounding site 4 was clean enough for individuals to walk right past.

Site 5 was the worst maintained of all the sites. This was the only bin that was brimming with garbage all seven days and was surrounded by trash. In at least two cases (sites 3 and 5), I observed individuals sorting through the trash and removing valuable recyclables. In at least three cases, I observed CMM staff actively picking up the trash surrounding the containers. In conclusion, it was evident that the CMM was providing fair trash service to this area. Although none of the containers appeared empty, trash levels within most containers appeared to remain the same, which suggested that the CMM did a good job emptying the containers and keeping the areas around the containers free from litter. This fDFWLQGLFDWHGWKDWDOWKRXJKWKH&00¶VVHUYLFHZDVDSSDUHQWDQGFRQVLVWHQWLQPDLQWDLQLQJ the area around at least 4 of 5 containers clean, it did not have the capacity to keep up with trash flow.

Lastly, the appearance of site 5 indicates that this container was perhaps not included along the scheduled route for service.

In conclusion, the CMM provided regular, consistent trash service. In limited instances, the service did not keep up with the production of trash. Trash levels in both observed districts remained relatively level. This indicates that the CMM is unable to keep up with the level of trash produced by residents and businesses to maintain empty bins. This was consistent with what we know about the

&00¶VWHFKQLFDOFDSDFLW\WRSURYLGHWKHVHUvice. The CMM was responsive with street sweepers as evidenced by their presence in both the Coop and Mafalala areas during the observation times. When trash surrounded the containers, the CMM acted quickly in dispatching workers to remove it. The

CMM reported that the weekends are the peak times for trash production, since pay day usually falls on

D)ULGD\DQGSHRSOHVSHQGPRUHWLPHDQGPRQH\IRUHQWHUWDLQPHQW7KLVUHVXOWHGLQD³UHFRYHU\´SHULRG of two days or so, in which the CMM had higher than normal levels of trash to pick up on Monday and

Tuesday. This might explain why there is always trash present in most of the containers; because the service could not keep up with the weekend deluge.

Beira

I conducted a trash observation experiment in two districts, which included the neighborhoods in which the focus groups were conducted. The purpose of this observation experiment was to monitor in a systematic way over a pre-GHILQHGDUHDWKH&0%¶VWUDVKFROOHFWLRQSLFN-up service along two major thoroughfares (Mateus Sans o Muthemba Avenue and Balazar Rebelo de Sousa Road) in a formal and informal neighborhood and determine if, indeed, the service was being conducted and compare the quality of the service. The units of observation were a total of 10 trash containers, 5 in each neighborhood. Because service in these two districts is daily, we chose the seven-day period and the same time every day to observe the volume of trash in and around these containers.

The first observation route had five sites in zone 1, Ponta Gêa, from Mouzinho de Albuquerque

Road moving west to the its intersection with Vasco Fernandes Homen Road, proceeding south to Dom

Francisco Barreto Road and then southeast to Mateus Sans o Muthemba Avenue, along the bay. The second observation route was also in zone 1, in the Inhamodima portion of Chipangara, which borders

Ponta Gêa and involved five sites. The observation route began at the turn-off of a paved road, Baltazar

Ribelo de Sousa, south onto the unpaved Vasco da Gama Road. From these, we moved east to Roberto

Ivens Road, to Road 1044 and Road 1071, all unpaved. We continued east onto a paved road, Capitães de Sena Road until it intersected with the paved Diogo de Couto Road, also known as the Mortuary

Road. These experiments were conducted over a period of seven days at the same time everyday.

The main finding was that, unlike Maputo, where trash collection service could not keep up

ZLWKWKHSURGXFWLRQLQWKHIRUPDOQHLJKERUKRRG%HLUD¶VWUDVKFROOHFWLRQLQERWKWKHIRUPDODQGLQIRUPDO neighborhoods appeared consistent with the production of trash. Unlike Maputo, where the contractors conducted some of the trash collection, the CMB provided all services directly.

Ponta Gêa

We examined 5 trash containers: Mouzinho de Albuquerque and the Police Station (site 1),

Mouzinho de Albuequerque and Vasco Fernandez Homem (site 2), Rua Dom Francisco Barreto and the market (site 3), Dom Francisco Barreto (site 4), and Mateus Sans O Muthemba (site 5). Overall, although service was consistent, it appeared that some containers received more regular service than others. With two exceptions, three of the containers never filled. Although the CMB provides trash service everyday, this neighborhood did not appear to receive service every day. The first day of the experiment was a Monday, when trash was pretty light. The pattern of trash accumulation indicated that pick up was likely twice in the seven-day period. Trash accumulated throughout the week and was cleared by day 5, which was a Friday. In one instance, the CMB removed a container for two days and trash began to accumulate in the spot where the container had once been. When the container was replaced, trash continued to accumulate in the pile next to the container. Unlike Maputo, where trash was a combination of paper and plastic products, trash was mostly organic (branches, leaves, and sand).

Site 1 had a steady increase in trash, starting Monday (day 1) 1/4th full and by day Friday (day

5) reaching 4/5ths full before the trash level reducing to 1/3. Sites 2 and 3 had a similar pattern of trash accumulation, being full by day 5. The site 4 container was removed day 2, a common practice in

Maputo and Matola, when containers are taken to the dump and emptied there instead of emptied into a trash compacting truck at the site. The container is usually replaced the same day or shortly thereafter.

In this instance, a period of two days passed when the CMB replaced the container, it did not clean the

SLOHRIJDUEDJHZKLFKKDGDFFXPXODWHGLQWKHFRQWDLQHU¶VDEVHQFH6LWHZDVDWUDVKGUXPDORQJPDMRU

road, which did fill quickly, because of its size, and appeared to be serviced more, because its trash levels varied on a daily basis.

Chipangara (Inhamodima)

Like Ponta Gêa, I selected five observation points along a route that included both paved and

XQSDYHGURDGVZLWKLQ&KLSDQJDUD¶V,QKDPRGLPDQHLJKERUKRRd. Observation began at the intersection of Vasco da Gala and Baltazar Ribelo de Sousa, a chapa stop (site 1), continued to Roberto Ivens 168

(site 2), to the intersection of two unpaved roads, number 1004 and number 1071 (site 3), east to the

Mortuguary road (site 4) and then ended at a trash drum on Capitães de Sena (site 5). Unlike Maputo, where trash accumulation outpaced collection in the formal neighborhood in comparison to the informal neighborhood, leading me to conclude that collection was actually better in the informal neighborhood, trash collection services appears to be more or less the same, with two pick ups per seven day period in both neighborhoods.

The pattern of trash accumulation in Inhamodima is similar to that of Ponta Gêa, with pick up twice a week and steady accumulation of trash throughout the week. The first day of observation was a

Monday, with the last day being a Sunday. Trash was highest on Monday and Sunday, with a pick-up likely taking place Monday afternoon and Saturday, based on trash accumulation patterns. Four of the five sites had high levels of trash on Monday, day 1, and were full or almost full by day 7, a Sunday, both the likely accumulation of weekend trash. Like Ponta Gêa, trash was a combination of organic waste (leaves, branches) and inorganic waste (plastic).

In conclusion, the patterns of trash accumulation in both Inhamodima and Ponta Gêa prove that

WKH&0%¶VWUDVKFROOHFWLRQVHUYLFHFDQNHHSXSZLWKWKHWUDVKSURGXFWLRQ$OWKRXJKWKH&0%FRXOG improve its service in the area of sweeping, as evidenced by the one container in Ponta Gêa, the trash in all ten observation sites never exceeded the containers. Trash accumulated at a steady pace throughout the week and the CMB, at least in the observation areas, was able to provide twice-weekly service that

often kept the container and surrounding area clean. All things being equal, the CMB is able to provide consistent trash service that keeps up with the production whereas Maputo struggles, especially in its formal neighborhoods, to keep up with trash production.

Matola

Matola has three Urban Districts: Matola Center (1), Machava (2), and Infulene (3). My two observation sites, Matola A and Matola C, are sub-bairros in Urban District 1, where we conducted two focus groups. Although adjacent, Matola A is the formal neighborhood and Matola C the informal neighborhood. I selected a total of three sites in Matola A and two sites in Matola C. Observation in

Matola A took place along two major streets (Nelson Mandela Avenue and Namaacha Avenue). Two sites were similarly selected along Salinas Road and Road 12026 in Matola C, which was a combination of paved and unpaved roads. Because service in these two districts is weekly, the observation period was between three to four weeks, one day a week (Thursdays in Matola C and Mondays in Matola A) to observe the volume of trash in and around these containers. Like Maputo, trash was mostly inorganic

(plastics, cardboard, paper) with some inorganic trash, specifically in Matola A, the formal neighborhood.

Although trash service in Matola A is regular, the collection is incomplete, leaving piles of trash next to the containers. Of the three sites that were studied over a three to four week period, all containers were serviced regularly. Two of the three containers were regularly overflowing by the end of the week. However, the CMMa did empty the containers. The main issue with the service was that garbage regularly accumulated or spilled around all three containers that continued to accumulate to the point that passing individuals would throw garbage on the piles even though the containers were empty.

Although one container was consistently empty, debris gathered around it. In the remaining two instances, the containers still had a substantial amount of remaining garbage (1/4 to 3/4ths full).

We observed two sites over three weeks in Matola C. Unlike Matola A, there were no containers for trash pick-up but rather individuals gathered their trash, mostly organic (branches, leaves and sand swept from the courtyard) and left it in large sacks, which the tractor team then collected.

Like Matola A, service was regular. Unlike Matola A, Matola C observation sites had more organic trash than inorganic, which is likely the result of consumption being lower, objects, such as plastic and cardboard, being re-used and all other trash being burned in the backyard. Although parts of Matola C were indeed inaccessible by even the tractor, they still appeared fairly clean. I observed people sweeping their backyards and burning trash.

Dondo

Dondo is the smallest in both population and square kilometers of the four case studies. It is divided into ten neighborhoods of which one, Central, is the subject of this observation exercise. I chose Central (formal neighborhood) and Macharote (Campo Estado section, informal). Central is completely paved whereas Campo Estado is not. I selected four sites in Central, on two parallel paved

VWUHHWVWKDWVXUURXQG'RQGR¶VPXQLFLSDOVWDIIEXLOGLQJDQGLWVQHZFRQIHUHQFHFHQWHUZKHUHWKH

Municipal Assembly and the Mayor meet. Taking its name from the soccer field, I observed five sites in Campo Estado, which encircle the actual soccer field. A small, informal market sprung around

Campo Estado, which assumed this name and the 5 sites are between the market stalls and the walls of the soccer field. Trash service occurs during the work week only (five days, Monday through Friday) and so the observation experiment was daily for five days. Like Matola and Beira, although service is daily throughout the city, service rotates from neighborhood to neighborhood in Dondo. The service in

Dondo in both these neighborhoods was consistent and the CMD was able to keep up with the flow of trash in the Central neighborhood. Although I observed changes in the levels of trash in the five observation sites in Campo Estado, there was never a complete elimination of trash, which was likely due to the lack of containers and difficulty in gathering all the trash. Furthermore, the piles that were closest to the main unpaved road in Campo Estado were picked up as the trash tractor has access but, in

the interior part of the bairro, residents often resort to burning trash. In both neighborhoods, I observed the trash tractor collecting garbage from the containers/sites of accumulation twice in one week although, as I discuss below, in Campo Estado, the actual collection occurred nearest to the main unpaved road.

In Central, the streets were paved. The CMD used the bottom half of large drums, which were numbered and remained in the same location throughout the week. Day one of the observation experiment was a Tuesday and trash in three of the four sites was understandably full to overflowing because of accumulation from a weekend without service. The trash was mostly organic (leaves and branches) due to the fact that most individuals sweep their backyards and the area in front of their homes daily and then dispose of these organic materials in the container. Interestingly enough, for an area of high population density, there was more organic trash than inorganic waste, unlike the Coop neighborhood in Maputo. Trash continued to accumulate at all sites on day 2 and by Thursday, day 3, the trash tractor, with a team of five men, emptied the containers onto a flatbed. By day 4, trash had begun to accumulate and by day 5, the following Monday, the tractor appeared to have passed again as all four containers were empty.

The same tractor that services the Central neighborhood also collected the trash in Campo

Estado, an adjacent informal bairro. I observed the tractor collecting trash twice, on Friday and on

Monday. Of the five sites, only two piles of trash appeared to have been picked up by the trash crew with the remaining three piles either continuing to accumulate or be burnt to dispose of accumulated organic trash. The two sites in which trash had been picked up were both adjacent to the main unpaved road in Campo Estado, with the remaining three piles in the interior of the bairro. Trash was primarily organic, with points of concentration along the walls of the soccer stadium, in areas where there was a concentration of merchants or near latrines.

In conclusion, in all four municipalities, we observed regular trash service and service was limited in informal neighborhoods because of the lack of access to the interior. However, this service

and extent of access to these informal neighborhoods varied. Maputo remains a unique case. It is in the only municipality which contracts service to both private firms in certain high-volume areas and microenterprises in informal and periurban neighborhoods. It is also the only municipality in which the observation sites received daily service, a hallmark of modern trash collection service. Although construction for a new paved roaGWRWKHLQWHULRURI%HLUD¶V0XQKDYDQHLJKERUKRRGLVXQGHUZD\

Maputo is the only municipality in which the informal neighborhood (Mafalala) had some paved streets, which increases access to these areas. However, in both Maputo and Matola, the formal neighborhoods had regular, visible service but this service could not keep up with trash production, as evidenced by the lingering piles of trash in Matola A and overflowing trash bins in one of the Maputo District 1 sites.

Interior portions of informal neighborhoods continue to experience a lack of regular service in

DOOIRXUPXQLFLSDOLWLHVEXWWKHOHYHORIVHUYLFHYDULHVIURPFLW\WRFLW\,Q%HLUD¶V,QKDPRGLPD neighborhood, the Secretary of the bairro calls the CMB regularly to have trash picked up from some interior points. In Dondo, the use of a tractor allows the CMD to pick up trash on unpaved roads, even if the narrower paths remain inaccessible. In Maputo and Beira, the municipality has started paving wider, interior roads in informal neighborhoods, which will eventually allow greater access to these areas. All four cities use tractors in some of the interior areas to pick up trash.

Based on the evidence from this experiment, not only does variation exist but we are able to ascertain some initial patterns of municipal performance. Based on this observation exercise, Maputo ranks as a middle performer because, while it was unable to keep up with the flow of trash, Maputo has done the most consistent collection in informal neighborhoods than the other cases. Beira and Dondo are high performers because they provide consistent service, which keeps up with the flow of trash in accessible areas. Lastly, Matola ranks as a low performer because the CMMa is unable to keep up with the flow of trash in both in its formal neighborhood.

Conclusion

Both experiments tell us that our dependent variable²municipal performance²varies in the

IRXUFDVHV0DSXWRLVUDQNHGDVDKLJKSHUIRUPHUEHFDXVH0DSXWR¶VDPEXODWRU\YHQGRUOLFHQVHSURFHVV was the shortest, most transparent and most professional and its trash collection is consistent and

FRPSUHKHQVLYHGHVSLWHWKHFRQWUDFWRU¶VLQDELOLW\WRNHHSXSZLWKWUDVKIORZLQWZRVLWHV'HVSLWHLWV good performance in trash collection, Beira is ranked as a middle performer because its license application process was circuitous. Dondo is a high performer in the area of trash collection, as

GLVFXVVHGDERYH)LQDOO\0DWROD¶VPXQLFLSDOSHUIRUPDQFHEDVHGRQWKHVHH[HUFLVHVLVUDQNHGDVORZ because it had the longest and most expensive license application process and not only could it not keep up with the flow of trash in Matola A, the trash consistently accumulated around the containers in all three observation sites (see Table 3.2). Why is there variation? Initial evidencHLQGLFDWHVWKDW0DSXWR¶V performance might be greatly affected by participation. Strong mayors in Maputo, Beira and Dondo might explain why performance is high in those municipalities whereas lack of consistent leadership and low per capita investment could be the causes of low performance in Matola. The following chapter isolates and examines how civil society in these four cases affects municipal performance of trash collection and vendor services.

Table 3.1. Expected versus Actual Performance

Civil Revenue Mayor Expected Actual Performance Society Party (1998 Performance to 2003 mandate)

Trash Vendor License Collection

Maputo High High FRELIMO : High Medium High Mayor switched, party same.

Beira Low High RENAMO: High High Medium Mayor switched, party switched.

Dondo High High FRELIMO: High High -- Same mayor, same party.

Matola Low Low FRELIMO: Low Low Low Same mayor, same party.

Table 3.2. Ambulatory License Process

Cost (USD) Steps Time

Maputo $45.28 6 22 days

Beira $25.32 8 38 days

Dondo - - -

Matola $58.49 8 55 days

,9(67$%/,6+,1*&,9,&3$57,&,3$7,21¶6

EFFECTS ON MUNICIPAL PERFORMANCE

Democracy - in particular, a healthy liberal democracy - also requires a public that is organized for democracy, socialized to its norms and values, and committed not just to its myriad narrow LQWHUHVWVEXWWRODUJHUFRPPRQ³FLYLF´HQGV6XFKDFLYLFSXEOLFLVRQO\SRVVLEOHZLWKD YLEUDQW³FLYLOVRFLHW\´ Larry Diamond (Diamond 1997, 5)

The concept of democracy as it is understood and elaborated in the West will encounter problems of application as does a pot of flowers replanted without considering climatic conditions or the environment. 65

Mozambican Jamisse Taimo (Thompson 1999, 195)

Introduction

The purpose of this chapter is to explore how participation in the four case studies affected municipal performance. I hypothesized that strong civil society would pressure a municipality to improve performance in the areas of trash collection and market services. I measured the strength of civil society by examining two factors: its access and control to resources and influence it had over municipal policy. While I found that strong civil society was not a pre-requisite for strong municipal performance, three high performing municipalities²Maputo, Beira and Dondo² provided examples of participation that facilitated improvements in municipal service. Civil society was fluid and adaptive to resource availability and that the type of donor assistance and relationship with the municipality were greater determinants of how civil society positively affected municipal performance than its strength. In Maputo, the voice of informal vendors was a strong lobby that orchestrated boycotts to change municipal policy. However, while it did influence municipal policy, its action did not necessarily lead to better service outcomes for informal vendors. In this

same case, street protests and a boycott against the imposition of a flat trash fee without consultation with residents resulted in a modified fee schedule based on energy consumption has been much better received. In Beira and Dondo, civil society extended municipal services without influencing municipal policy. However, in Dondo,, a donor-funded participatory governance process allowed residents to influence the types of infrastructure projects the municipality funded.

How civil society and the state interact is a classic and timely area of study. As a post- colonial, post-socialist, new democracy, the shape and function of civil society takes a form

VSHFLILFWRWRGD\¶VMozambique. Decentralized municipalities receive donor assistance to encourage citizen participation processes or organizations that collect trash. Simultaneously, municipalities work with inherited colonial associations that keep order in markets as an extension of the state. Mozambique is grappling with the perennial question of state building in which civil

VRFLHW\LVERWK³WKHUHDOPLQZKLFKWKHH[LVWLQJVRFLDORUGHULVJURXQGHGDQGLWFDQDOVREHWKH

UHDOPLQZKLFKDQHZVRFLDORUGHUFDQEHIRXQGHG´ (Cox 1999: 4). As discussed in Chapter 2, to be able to have a complete view of civil society in Mozambique in order to analyze how it affected municipal performance, I drew from scholars who provided insight into the way civil society does

(and does not) increase citizen voice and participation by interacting with the democratizing, post- socialist Mozambican state. My focus in this chapter is not the study of Mozambican civil society but rather how civil society organizations I studied in each of the four cases affected the

PXQLFLSDOLW\¶VDELOLW\WRSHUIRUPWUDVKFROOHFWLRQDQGPDQDJHPDUNHWV

I found that a strong civil society is not a pre-requisite for high municipal performance; however, three of the four cases²namely, Maputo, Dondo and Beira²had degrees of citizen participation that improve municipal performance. In one case, a civil society organization²an informal vendors association in Maputo²pressured the municipality to concede municipal resources and responsibilities to it, acting as an independent political force and providing some voice to informal vendors, who boycotted fee payment to change municipal behavior as will be discussed below. In Maputo, I also found neighborhood associations that began collecting trash,

which, with the help of donors who joined them in lobbying the CMM as an independent political force, became private microenterprises with city contracts. Citizen protests against the introduction of a flat trash fee also provided evidence of the efficacy of voice in improving municipal service. Lastly, in Dondo, donors fostered citizen participation through a donor-funded participatory budget processes that become an institutionalized municipal feedback mechanism,

KHDGHGE\WKH&0'¶VWRZQFRXQFLORUVZKRUHSRUWHGEDFNWRWKe Mayor, in an instance where civil society organizations were weak. This mechanism increased citizen participation in informing decisions the CMD made on infrastructure investments and required town councilors and the mayor to justify these publicly in regular public meetings.

In all four municipalities, I observed that Merchant Associations were extensions of the state, actively in performing municipal functions on behalf of the municipalities either with delegated municipal authority, funding concessions or both (see Table 4.2). However, I found in

Maputo, Merchant Associations were strong and had greater voice in how the market was managed because the CMM conceded a percentage of what they collected in vendors fees to these

Associations to use for market improvements of their choosing. In trash collection, I found that groups changed over time to become self-sustaining institutionalized service providers, extending the reach of the municipality, such as the trash microenterprises and neighborhood groups in

Maputo. Also, in Beria, Maputo and Dondo, aid played a critical, capacity-building role that contributed to municipal service delivery. In other instances, civil society organizations contracted in some municipalities until consistently activated by donor funding, such as neighborhood groups in Beira, which did not improve municipal service in the long term but did provide sporadic neighborhood cleanup in the short-term, absent support. In Maputo, Beira and Dondo, civil society was either strong or active due to access to (and control over) resources and their connection to the municipality. In Matola, performance was low. Merchant Associations were weak because they lacked both access to resources, a connection to the municipality, and existed in the presence of

corruption and municipal management. In trash collection, there were no organizations working in this area.

This literature enabled me to H[DPLQHKRZSURFHVVHVOLNH'RQGR¶VSDUWLFLSDWRU\EXGJHW process, and organizations, such as Merchant Associations and neighborhood trash organizations, affected municipal performance by allowing me to explore how this participation created (or did not create) pressure to improve these services. In sum, I found that participation was present in three of the high performing cases in varying degrees and that strong civil society organizations were not a prerequisite to strong municipal performance. Civil society organizations are fluid and adaptive to resource availability. They affect municipal performance based on how financially sustainable they are and the degree to which municipalities rely on them to act on their behalf.

Lastly, donors in two cases²Maputo and Dondo²helped increase participation by providing platforms for civil society to pressure these municipalities.

Mozambican Civil Society: Current State and Departure Point

Key findings of recent studies into Mozambican civil society and citizen participation indicate that it is weak for two reasons: first, organizations lack financial and organizational capacity; and second, citizens in practice have little economic and political influence. A 2007 study sponsored by the Mozambican Foundation for Community Development and the United Nations

Development Program found that civil society has difficulty effectively monitoring the state

EHFDXVHWKHLQWHUDFWLRQEHWZHHQWKHWZRLV³VXSHUILFLDODQGLQIRUPDO´GXHWR³SXEOLFDQGSULYDWH institutions (that) are still very underdeveloped, rudimentary, sharply focused around bipolarized

SROLWLFDOSRZHUDQGZLWKRXWPHFKDQLVPVIRUWKHHIIHFWLYHUHSUHVHQWDWLYHQHVVRIFLWL]HQVKLS´

(Francisco, Mucavele, Monjane, Seuane 2008, xxvi). Donors tend to focus on civil society

100

organizations in urban areas, missing opportunities to work with rural organizations.66 Aside from

IDFWRUVOLNHFLYLOVRFLHW\FDSDFLW\DQGSROLWLFDOFRQVWUDLQWVWKHODFNRI³FULWLFDOFLWL]HQVKLS´LV

another factor that hampers civil society in Mozambique. Mattes and Shenga concluded in a

national study of Mozambican political attitudes that while participation in organizations

(specifically, community development organizations) does enable people to form opinions about

the performance of democracy and governance, Mozambicans are less critical of their

JRYHUQPHQW¶VSHUIRUPDQFHWKDQRWKHU$Irobarometer countries because of a low-level of formal

education, political information and interest in politics (Mattes and Shenga 2007).

As previously stated, I set out to investigate how participation created pressure on the

municipalities to improve service provisions. Specifically, I hypothesized that civil society

organizations that are strong can pressure municipalities to improve performance in the two test

areas (see Table 4.1). I also looked for other evidence of participation such as the ability for

citizens to voice concern over and affect municipal policy. I measured the strength of civil society

organizations in two ways: access and control to resources²RUZKDW,WHUP³ILQDQFLDO

VXVWDLQDELOLW\´²and influence over municipal policy. I measured financial sustainability by

examining if civil society organizations have access to regular income over which they have

discretion. Maputo was the only case, in which Merchant Associations, the informal merchants

lobby, and the neighborhood trash collection organizations had access to regular income over which

66 For the purposes of initial discussion in Chapter Three, I used the Directory of Development Organizations to get a sense of where civil society organizations were located. However, I rely on the FDC/UNDP report for this chapter since it is a thorough study of Mozambican civil society. The FDC/UNDP report uses a sample of 477 civil society organizations from a census of 5,000 non-profit organizations done in 2003 by the National Institute of Statistics, in addition to interviewing a statistically representative sample of 4,015 people. While the FDC/UNDP report conclusions differ with the number of civil society organizations in the Directory on the degree to which civil society organizations are located in the capital, both sources reflect that the plurality of civil society organizations are located in the capital (90 percent for the former and 21.31 percent in the later). FDC Mozambican Civil Society Within: Evaluation, Challenges, Opportunities and Action, 2007 and Directory of Development Organizations: Guide to International Organizations, Governments, Private Sector Development Agencies, Civil Society, Universities, Grantmakers, Banks, Microfinance Institutions and Development Consulting Firms, Edition 2008, Volume I.B/Africa, Mozambique, Resource Guide to Development Organizations and the Internet, http:/www.devdir.org.

101

they had some discretion and where I observed both the informal vendors association and neighborhood trash collection association independent political forces that lobbied for access to, and received shares of, municipal revenue and changes in municipal policy. Maputo was also the only case in which I observed Merchant Associations²as extensions of the state²receive a share of municipal income for the maintenance of their respective markets. A separate, but relevant, finding was that donors in Maputo worked with neighborhood associations to both build their capacity to collect trash in informal neighborhoods and lobby the municipality for contracts to create a steady stream of income. I measured how civil society organizations affected municipal performance by examining episodes in which civil society organizations successfully sought and received concessions in municipal policy.

I also found in two cases that participation manifested itself in more direct citizen participation in the form of boycotts and neighborhood hearings that affected municipal decision- making. In Maputo, 2002 citizen protests boycotting a new trash fee caused the municipality to alter its policy to a more rationale fee schedule. In Dondo, neighborhood hearings in the form of the

PXQLFLSDOLW\¶VSDUWLFLSDWRU\EXGJHWSURFHVVSURYLGHDFKDQQHORIGLUHFWFRPPXQLFDWLRQEHWZHHQ citizens and the municipality related to infrastructure investments and forced the municipality to justify these investment decisions on a regular basis. In both instances, donors worked with the municipality to resolve boycotts in Maputo or institute citizen-municipal channels as was the case in Dondo.

In summary, based on my initial review of performance in Chapter 3 and theory in Chapter

2, I hypothesized that Maputo, Beira, and Dondo would have strong civil society organizations in the two test areas and Matola to have weak civil society organizations (see Table 4.2). I found that financial sustainability and influence over municipal policy²two of my measures of civil society strength²were not pre-requisites to high municipal performance in the two test areas. The type of donor assistance and the concessions the municipality made to delegate municipal responsibilities to those organizations were two factors present not only in Maputo but also in Dondo and Beira that

102

led active civil society organizations to facilitate access to municipal services even though these

organizations were not strong, as was the case in Dondo and Beira. I also found evidence of how

citizen voice through protests in Maputo effectively influenced the way the CMM charge these

fees.

In the following sections, I first review briefly how municipal market management and

trash collection services function. Then, I examine the participation in each case that work in these

two areas and how they affect municipal performance of those two services. Lastly, I conclude the

section with a summary of findings.

How Services Work

Merchant Associations General Organization

As discussed in the introduction, market management was designated as a municipal

responsibility as a part of decentralization reforms in 1998.67 Unlike trash collection, for which the

city bears full responsibility, the municipality delegates daily market management to the Merchant

Associations. Merchant Associations existed in the time of Portuguese colonization and were

designed to keep order in the markets. These bodies are governed by municipal regulations. How

these bodies were governed²specifically, how municipalities chose to apply options for revenue

sharing, elections of members, and fee-schedules²incentivized the degree to which Merchant

Associations felt ownership over market improvements. I discuss this in each case below.

67 Market management was a function of local government even before decentralization. The major distinction is that under decentralization, municipalities became politically autonomous with an extended degree of fiscal autonomy, incentivizing mayors to examine how to encourage fee-generating activities. Teodoro Andrade Waty, Autarquias Locais: Legislação Fundamental, W&W Editora, Maputo: 2000: 111.

103

Merchant Associations are typically comprised of vendors and headed by an elected President,

Vice President, Treasurer and Secretary.68 These associations are neither purely outside the state

nor purely a part of it. They are charged by the municipality with maintaining order and sanitary

conditions in the markets and compelling vendors to pay municipal fee collectors, all of which are

municipal responsibilities. Although all markets have municipal staff in the form of a market chief

and fee collectors, in order to compel vendors to pay, the municipality relies on the Merchant

Associations leadership because they often know the vendors well, especially in the large markets.

7KH$VVRFLDWLRQ¶VOHDGHUVKLSERWKFDPSDLJQHGWREHHOHFWHGDQGWKHPVHlves are vendors; therefore,

they are seen as more credible. Vendors often see market leadership as the members of the

Merchant Association, not the municipal staff, and call upon the Merchant Association to resolve

vendor disputes. Merchant Associations negotiate between representing vendor interests and

responding to the municipality by imposing discipline in the market.

68 The Conselho Municipal de Dondo (CMD) classifies the markets and Merchants Associations differently than the other three municipalities. This will be further discussed in the Dondo section.

104

Solid Waste Management (SWM) General Organization69

Trash collection is another municipal responsibility as per 1998 decentralization reforms (Waty

2000, 111). The municipality is responsible for nine general functions surrounding trash collection,

which include provision of public containers and collection and disposal of trash.70 Case study

cities provide this service in a variety of ways, which range from private contractors, city

employees, and microenterprises, and at a variety of levels from modern trash trucks and skip

loaders to tractors and handcarts. Unlike Merchant Associations, which are an established set of

organizations that balance vendor and municipal interests, in the area of trash collection,

neighborhood organizations sprouted up largely as a result of donor funding. Organizations take

difference shapes, with some comprised of neighborhood volunteers while others become

microenterprises. Unlike market management, where it is labor intensive to collect fees,

69 Both national and local law provide for the provision of trash collection for residents of municipalities, which is rooted in the care and maintenance of the environment. According to the Constitution of the Republic, article 90:

(1) All cities have a right to live in an balance environment and have the right to defend it;

(2) The State and the local autarchies in collaboration with associations that defend the environment adopt policies of the defense of the environment and see to the rational utilization of all natural resources. (Constitution of the Republic of Moçambique, Plural Editores: Maputo, 2007, p. 25)

The National Environmental Policy, approved by the National Assembly in Resolution no. 595 of the 3rd of August, indicates that there is a interdependence between development and the environment.through the Ministério da Coordenação Ambiental (MICOA) and the municipalities. MICOA has produced a series of regulations governing solid waste management, which include the Strategy of Integrated Solid Waste Management, February 2004 and Decree no 13/2006 of the 15 of June. One MICOA decree, 13/2006 of 15 of June specifically vests municipalities with the responsibility of approving the standard of solid waste management (SWM), fixing the fees for the provision of service, removal and depositing of said waste and licensing establishments that produe this waste. These responsibilities are further codified in Law no. 11/97 of the 31 of May of 1997, article 25, subsection b, that the municipality collect and treat trash and article 13, which allows municipalities to charge fees (Andrade, Autarquias Locais: Legislação Fundamental: 111 and 105). Law no. 11/97, articles 70 and 71 specifically provide for the charging the collection and treatment of solid waste and the application of fines for individuals who violate municipal regulations of no more than 10 times the minimum national wage (Andrade, 133). Each municipality is charged with enacting its own regulation governing municipal activities, which include SWM.

70 Responsibilities typically include location of public containers for disposal of trash, trash pick-up, transportation of trash, storage of equipment and vehicles related to SWM, transfer for trash treatment of solid waste, sweeping, and elimination of solid waste and final internment, typically in an open-air dump site..

105

municipalities have arrangements with Electricidade de Moçambique (EDM) to collect trash fees

through electricity bills. EDM charges each municipality a monthly flat fee to perform this service,

ZKLFKLVDSHUFHQWDJHRIWKHWRWDOWUDVKIHHVFROOHFWHGWKDWYDULHVE\PXQLFLSDOLW\0DWROD¶V

SHUFHQWDJHZDVWKHORZHVWDWSHUFHQWZKLOH%HLUD¶VSHUFHQWDJHZDVWKHKLJKHVWDWSHUFHQW

EDM not only has the ability to collect fees but also the ability to apply an effective sanction for

unpaid fees by cutting off the electricity. This sanction is, of course, only effective for those who

have electricity, which the majority of Mozambicans do not (EDM 2011).71

Analysis by Case

Maputo

I hypothesized that strong civil society organizations influenced the high level of

performance in the two test areas in Maputo (CMM) (see Table 4.2). Specifically, I expected that

civil society organizations would be able to exert pressure over the municipality to manage markets

and collect trash. For market management, I found that Merchant Associations were both

financially sustainable and influence the municipality to respond to their requests. I also found that

a strong informal vendors association²ASSOTSI²was financially sustainable and pressured the

PXQLFLSDOLW\WRFKDQJHLWVSROLFLHVRQIHHFROOHFWLRQLQ$66276,¶VIDYRU+RZHYHULWGLGVRLQ

order to try to supplant the CMM in informal markets rather than to pressure the CMM to provide

better services in order to preserve access to its membership fee revenue. For trash collection, I

found neighborhood organizations that lobbied the municipality for service contracts to extend

service to informal neighborhoods. So, they became both financially sustainable and municipal

service providers.

71 EDM covers only 36 percent of Mozambique. The south, where Maputo and Matola are located, has the highest level of coverage at 43 percent. The center, where Dondo and Beira are located, has the lowest at 28 percent.

106

In terms of financial sustainability, CMM Merchant Associations received a monthly payment of 10 percent of the fees generated in that market from the CMM, which they managed and used towards new bathrooms, water fountains, guards and cleaning services. These groups were the only ones of the cases in which the municipality conceded 10 percent of vendor fees to vendors. In the area of trash collection, I found that neighborhood microenterprises, which started as associations funded by donors, grew to pressure the municipality to provide contracts, which

WKHQFRQYHUWHGWKHPLQWRSULYDWHFRQWUDFWRUVH[WHQGLQJWKH&00¶VUHDFKLQWRWKHLQIRUPDO

QHLJKERUKRRGV8QGHU'LDPRQG¶VGHILQLWLRQWhese organizations might be considered private² thereby disqualifying them as civil society organizations (Diamond 1993). However, these private contractors either started as neighborhood organizations that become microenterprises or continued to remain neighborhood organizations that donors developed to become contractors and fill a gap in service.

Merchant Associations

8QOLNHWKHRWKHUWKUHHFDVHV0DSXWR¶V0HUFKDQW$VVRFLDWLRQVwere strong because they were financially sustainable²or had a steady, reliable source of income they controlled²and they influenced the municipality to provide better services. This is due to two factors found in Maputo, which were not found in the other cases. First, the CMM incentivized Merchant Associations to perform by providing and enforcing institutional standards that govern the composition and function of the Merchant Associations. As stated in the introduction, the CMM installed institutional accountability²such as defined terms of Merchant Association leadership and fee schedules, which are generally the same across formal markets²that drive these Associations to perform. Second, the CMM gave them access to 10 percent of fee collections in order to gain their cooperation in managing the markets.

107

While Merchant Associations received only 10 percent of total vendor fees collected in their

markets from the CMM, this concession incentivized Association leadership to improve conditions

in the markets for vendors because they had some autonomy in deciding how to apply the funds.

For example, the Central Market Merchants Association used its 10 percent to pay its water bill and

was able to convince the CMM to assume some of the debt the market incurred in upgrading

infrastructure. As a result, the market ran a surplus because the Association collected more in

bathroom fees than it receives in the 10 percent pay-back.72 Although the Xipamanine Formal

Association ran a monthly deficit, the Association collected a steady stream of income from

bathroom fees, which equated to half of the 10 percent monthly pay-back. The president indicated

that the water pump and new sidewalks along the front of the formal market were provided by the

10 percent. He was trying to charge more bathroom fees to cover additional projects. In both

markets, the Merchant Associations have the technical capacity (i.e., track expenses, manage

projects) to administer the 10 percent pay-back and had strong working relationships with their

municipal counterparts.

Regular elections instituted by the CMM created an additional incentive for Merchant

Association leadership to be responsive to vendors and use funds judiciously. The Central Market

Merchants Association president himself was experienced, valued his campaign promises, and had

a good working relationship with the Market Chief. The president of the Xipamanine Formal

Market Merchant Association was also an experienced vendor who had a cooperative relationship

with the Market Chief. Both indicate that they were elected to fulfill campaign promises.

72 First, the CMM agreed to pay off market debt, the only market in the study with such an arrangement. Second, receipts from bathroom charges adequately cover almost all monthly charges, except for the water bill. The Association president indicated that they pay roughly $425 USD/month for guards, cleaning people, cleaning and drainage of the bathrooms and a professional accountant who balances the books monthly (the only market with such an arrangement). Receipts from the bathroom fees total anywhere between $555 USD to $629 USD per month. Third, the monthly water bill is covered by the roughly $482 USD/month from the 10 percent payback. Author interview, Central Market Merchants Association President, Maputo, October 1, 2008.

108

ASSOTSI

Of the four municipal cases, ASSOTSI in Maputo was the only strong civil society

RUJDQL]DWLRQLQWKHDUHDRIPDUNHWPDQDJHPHQWWKDWPHWWKHFULWHULDDVRXWOLQHGE\'LDPRQG¶V

GHILQLWLRQDVD³YROXQWDU\VHOI-generating, at least partially self-supporting (and) autonomous from

WKHVWDWH´ 'LDPRQG, 6). I classified ASSOTSI as strong because it was financially-

sustainable due to its access to member fees and, as the following episodes will demonstrate, they

were able to influence municipal policy in their favor. However, as the discussion will show,

despite being a strong civil society organization, ASSOTSI used its influence to preserve its own

power base rather than work with vendors to utilize the infrastructure the CMM was building and

pressuring the municipality to accommodate informal vendors.

Founded in 1999 to organize informal vendors, ASSOTSI represents informal vendor interests

and asserted itself fairly quickly within the informal markets, collecting fees and organizing.

ASSOTSI was seen by informal vendors as a much-needed advocate; however, this advocacy

FRPHVDWDSULFH7KHIROORZLQJIRXUHSLVRGHVGHPRQVWUDWHERWK$66276,¶VVWUHQJWKDQGKRZLW

was used to forward its own interests rather than pressure the CMM to be responsive to vendors.

The first example occurred in 2007. The interests of the CMM and informal vendors came to a

head. In an effort to increase revenue collection, the CMM enacted policies to formalize the

informal Merchant Associations and collect standard fees in the informal markets. However, the

CMM ran into competition. ASSOTSI charges vendors a 10-30 MTn/month membership fee,

which they reportedly use to reinvest in informal markets.73 However, the markets still have no

73 There exists some confusion about the vendor quota payments to ASSOSTI. Vendors, and the president of ASSOSTI himself, seem to have different responses to this question. Some indicate that it is 1 MTn per day, totaling 30 MTn/month and others indicate that it is 10 MTn per month. When I requested records of collections from the ASSOSTI president, which he should have since at least 2007 given that the CMM entered into an MOU with ASSOSTI that required them to give receipts of the payments they collected, I was told that they were not available. Amounts that ASSOSTI pays out to its affiliate Merchant Associations were also similarly unavailable. ASSOSTI then gives back an unspecified amount, which the Association can use for market maintenance. 109

potable water, public illumination, or water pump, stalls or walls. Some vendors see ASSOTSI

SD\PHQWVDQG&00PDUNHWIHHV³FRQVWLWXWLQJUREEHU\´DQGIHHOFRHUFHGLQWRPDNLQJWKHVH

³GRXEOH´SD\PHQWV 6DPER 74

Mayor Comiche wanted to end parallel payments to ASSOTSI because it was not clear how

YHQGRUVZHUHEHQHILWWLQJDQGLWFRQIRXQGHGWKH&00¶VDELOLW\WRFROOHFWIHHV2Q)HEUXDU\

Mayor Comiche met with ASSOTSI and shortly thereafter issued a service order, No 4/GP/2007, in

which it was prohibited for ASSOTSI to charge this fee (Ibid). However, ASSOTSI continued to

charge. Within one month, Mayor Comiche issued another memorandum, which overturned the

SUHYLRXVSURKLELWLRQ+HDFNQRZOHGJHG$66276,¶VULJKWWRFKDUJHLWVPHPEHUVYROXQWDU\³IHHV´

and outlined specific guidance to differentiate the ASSOTSI fee and the Directorate of Markets and

)DLUV '0) ¶VIHH

:KDWEURXJKWDERXWWKLVUDGLFDOVZLIWVKLIW"$FFRUGLQJWR$66276,¶VSUHVLGHQWKLV

relationship with President Guebuza enabled him to pressure the CMM to modify its policies. In

this instance, the ASSOSTI president intimated that he called the President to explain why this

prohibition was unnecessary and reminded him that ASSOTSI was the only organization that could

control informal vendors.75 The ASSOTSI president felt that the CMM unnecessarily interfered

when ASSOTSI had already organized the vendors. In response to the confusion over ASSOTSI

Whereas the CMM has the 10 percent pay-EDFN$66267,VWLOO³KHOSV´DFFRUGLQJWRSUHVLGHQW when the 10 percent is not enough. Previous to the 10 percent, Associations kept the collections but it is unclear how these collections were used. In Xiquelene Informal Market, if the Association made more PRQH\EHIRUHWKH&00¶VSUHVHQFHWKH\GLGQRWDSSHDUWRVSHQGLWRQPDUNHWLPSURYHPHQWVOLNHWKH bathrooms and offices, because this infrastructure was built with the 10 percent. This implies that what DMF staff has said has some truth ²WKH&00¶VGHFHQWUDOL]HGRYHUVLJKWRIWKHPDUNHWVKDVUHVXOWHGLQ a steady fee collection which returns to the market at least the 10 percent payback and more WUDQVSDUHQF\LQWKH0HUFKDQW$VVRFLDWLRQ¶VZRUNBoth Xiquelene and Xipamanine Association members assert that improvements are the result of the 10 percent and unspecified ASSOSTI funds.

74 2QHYHQGRUH[SUHVVHGWKDW³WKLVFRQVWLWXWHVUREEHU\EHFDXVHZHQHYHUEHQHILWWHGIURPDQ\DFWLYLWLHV that were the fUXLWRIWKRVHFKDUJHV´RQHPDUNHWFKLHILQGLFDWHGWKDWWKHnúcleos de base (informal Merchants Associations) were downright extortionary.

75 Author interview, ASSOTSI President, Maputo, September 30, 2008.

110

IHHV'U&RPLFKHVLJQHGDPHPRUDQGXPRIXQGHUVWDQGLQJZKLFKDFNQRZOHGJHG$66276,¶VULJKW

to charge its members fees, as long as they issued receipts that distinguished them from CMM fees.

ASSOTSI flexed its muscles in three additional episodes. In 2006, the CMM revised the

vendor fee schedule. However, vendors had difficulty making the payments because of

simultaneous devaluations of the metical.76 $66276,RUJDQL]HGD³ZRUNVWD\´LQZKLFKYHQGRUV

stopped paying fees, resulting in the CMM lowering fees. The CMM has since relied on ASSOTSI

to help compel vendors to pay even the lowered fees.

An earlier episode provides further evidence of how ASSOTSI exercises its influence to

compel municipal policy changes. Maputo experienced a cholera outbreak in 2003. As a result, the

CMM closed the restaurants in the markets while also requiring merchants to continue paying fees

for tKHWKUHHGD\VWKDWWKH\ZHUHFORVHG7KH$66276,SUHVLGHQWSXWWKH&00¶VWRZQFRXQFLORUV

and chief medical officer publicly challenged them to come down to the restaurants where the

outbreak supposedly took place. The medical officer ruled that cholera was not a problem and the

restaurants re-opened.77

Lastly, in an episode that took place in April 2008, ASSOTSI and the CMM clashed on the

number of ASSOTSI members with ASSOTSI arguing it had more members and the CMM arguing

it had less. The solution to the episode of the previous year was that ASSOTSI could charge its

members quotas as long as it identified these charges with its own receipts and only charged its

members, who should identify themselves with membership cards. However, ASSOSTI asserted

that it represented all informal vendors and so, they continued to charge vendors. ASSOTSI

SURSRVHGWKDWWKH&00SRVWGDLO\UHFHLSWVLQHYHU\PDUNHWDQGDOORZ$66276,³DVWKHODUJHVW

HPSOR\HULQWKH&LW\RI0DSXWR´WRSLORWFROOHFWLRQLQRQHPDUNHWDQGFRPSDUHLWWRWKH&00¶V

76 Maputo Municipal Assembly, Resolution 47/AM/2006 of the 29 September.

77 During focus group interviews in Mafalala, participants indeed mentioned the sickness and cholera outbreaks that occur in their neighborhood and, although they did not bring up this particular instance, they did express that barracas and street vendors are the cause of pollution and health problems in the bairro.

111

performance.78 The CMM rejected the notion of allowing another entity to manage the markets (Na

Cidade de Maputo 2011).79 Comiche further indicated that the CMM should launch a census of

legally recognized vendors so they could determine the number of ASSOTSI affiliates (Ibid). In

responVHWKH$66276,SUHVLGHQWDJDLQLQYROYHG3UHVLGHQW*XHEX]DZKRWDVNHGWKH&00¶V'0)

town councilor to establish a partnership between ASSOTSI and the CMM.

These episodes demonstrate how ASSOTSI used its influence to preserve its political standing

and access to member fees arguably at the expense of the formalization of service provision

through the CMM. ASSOTSI did lobby the CMM in ways that benefited informal vendors as

evidenced by the reopening of restaurants after the cholera outbreak and the work stay, which

revised fees. Although ASSOTSI established a practice of charging vendors and took credit for

organizing the markets and creating infrastructure (i.e., funding roofs, water and electricity), there

was little evidence of their role in these improvements. In both Xiquelene and Xipamanine

Informal Markets, the Association presidents mentioned that ASSOTSI provided them

supplemental funds for projects in the market although neither was able to provide me an example

or show a project that was paid fURPWKHVHIXQGV,WZDVWKH&00¶VSHUFHQWSD\EDFNWRZKLFK

they referred when they discussed improvements. When I asked for numbers or reports indicating

the flow of membership dues they collected and how they were used, the ASSOTSI president could

not provide them. He indicated that ASSOTSI provides a percentage of collected quotas back to

each market, but he could not show me proof or even give specific examples of amounts for the

major markets. The lack of proof of collections and ASSOSTI-related investments indicates that

funds collected through membership fees may not be utilized in the way ASSOSTI leadership

suggested.

78 Author interview, ASSOTSI President, Maputo, September 30, 2008.

79 7KHSUHVVUHOHDVHTXRWHV'U&RPLFKHDVVWDWLQJ³WKHHQWLW\WKDWPDQDJHVWKHPDUNHWVLVWKH&00 and in whichever other instances that might differ, the objective would have be to coordinate with the PXQLFLSDOLW\DQGWKH0XQLFLSDO$VVHPEOLHV´Na Cidade de Maputo: Taxas Dividem Município e Informais, April 25, 2008.

112

In conclusion, ASSOTSI was an example of a strong civil society organization because it was

financially sustainable and influenced municipal policy. Specifically, ASSOTSI had access to a

revenue base through its members. While I was unable to substantiate total estimates of member

fees, the episodes demonstrated that ASSOTSI organized and charged a large number of informal

vendors. 7KHHSLVRGHVDOVRVKRZHGWKDWLWVXFFHVVIXOO\FKDOOHQJHGWKH&00¶VFODLPDVWKH

authoritative presence in the informal markets. While strong, ASSOTSI used this strength to

preserve its access to member fees rather than improve access for informal vendors to municipal

services.

Trash Collection

,Q0DSXWRZDVNQRZQDVWKH³FLGDGHEDL[ROL[R´ WKHFLW\XQGHUJDUEDJH 80 By 2008,

the CMM provided daily trash collection in formal areas of the municipality and some service to

informal neighborhoods. How was the CMM able to implement these changes? A combination of

donor support of fledgling community organizations, citizen protest, and microenterprises and

municipal leadership explain the improved outcome. Specifically, I examined how citizen protests

in 2002 against the introduction of flat trash fee pressured the CMM to re-examined and re-

introduce an alternative, which was better received by Maputo residents. Also, I observed how civil

society in the form of two neighborhood organizations was strong. Specifically, I examined two

specific examples of organizations, which evolved into contractors for the municipality, thereby

H[WHQGLQJWKH&00¶VDELOLW\WRSURYLGHWUDVKFROOHFWLRQLQDSRUWLRQRILQIRUPDOQHLJKERUKRRGV

7KHVHRUJDQL]DWLRQVFRQWDLQHGFRPSRQHQWVRI'LDPRQG¶V³YROXQWDU\VHOI-generating, at least

partially self-supporting (and) autonomous from the staWH´EXWLQRQHFDVHWKHRUJDQL]DWLRQV

became a private microenterprise (Diamond 1997, 6). As contractors, these organizations became

80 Author interview, Town Councilor for Sanitation and Salubrity, Maputo City Hall, October 7, 2008. One long-time resident of an upscale neighborhood recalled that flies would regularly float into her kitchen²on the 12th floor²from the piles of trash on the streets.

113

financially sustainable and influenced municipal policy such that they both received a share of municipal resources and H[WHQGHGWKH&00¶VDELOLW\WRSURYLGHWUDVKFROOHFWLRQLQWRSUHYLRXVO\ inaccessible informal neighborhoods.

Protests, Organizations, and Donors:

Determining Revenue and Lobbying for Contracts To Fill in the Gaps

The 2000 floods created the worst flooding Maputo had seen in a century. The piles of trash that accumulated as a result of these floods and the lack of ability for the CMM to adequately respond created situations in which one local resident, living on the near-top floor of a 10-story building, recalls swarms of flies floating into the kitchen on a daily basis. In 2001, a private contractor compounded the problem by walking off the job further prompting the central

JRYHUQPHQW¶VLQWHUYHQWLRQLQSURYLGLQJHPHUJHQF\DVVLVWDQFHWRFRllect trash. By 2002,

GTZ/AGRESU began work with the CMM to introduce a flat trash fee.

A 20 MTn (less than $1USD) trash fee was passed by the Mayor and City Council in 2001 and went into effect the following year. The timing of introducing this fee could hardly have been worse. As citizens were being asked to pay for trash collection, piles of trash were at their highest and so citizens were outraged that they were now being asked to pay for a service they did see functioning. Citizens voiced their anger at City Council meetings. Residents in rich and poor sections of Maputo blocked major arteries, including one to the airport in what seemed to be a simultaneous outpouring of frustration among city residents at the lack of service and burden of a new fee. This fee was repealed in five months and the CMM did not collect any fees for the rest of the year. However, the CMM did roll out an education campaign regarding the role and relevance of charging a trash fee so residents could learn more about it.

Throughout this time, AGRESU worked closely with the CMM to do feasibility studies to understand the market for charging trash fees and how to design a fee scale that would be 114

compatible with long-term sustainable financing of trash collection while being equitable. By mid-

2003, these studies resulted in a fee scale that was linked to electricity usage that shifted the burden

of payment to those that, in theory, produced more trash and could likely afford to pay more in the

trash fee. The CMM, with AGRES8DOVREHJDQWRH[SORUHKRZWRFKDUJH³ELJ´SURGXFHUV LH

commercial producers) at a level consistent with the work the City did to collect their trash, thus

shifting costs in a more rational fashion. These protests and subsequent repeals and adjustments

also forced the CMM to work with neighborhood organizations in a new way²as contractors.

Today, Maputo has contractual relationships with over 20 microenterprises and community

organizations, which provide service to several informal neighborhoods. However, this was the

result of almost two years of negotiations after the 2000 floods between a small subset of

neighborhood organizations and the CMM, with the assistance of donors. The 2000 floods and

associated donor assistance combined to both create fertile conditions for the development and

sustenance of microenterprises and motivated the CMM to work with them in Urbanização,

Maxaquene A, and Xipamanine bairros.81 In 2000, Médicins San Frontéres (MSF) began cholera

prevention remediation in Urbanização to combat flood after-effects.82 A coalition of community

leaders formed, which became known as the ADASBU. At first, they volunteered to sweep paths

and then they turned into a business that hired community residents to collect trash door-to-door for

2MTn daily per home. MSF picked up the shortfall in fees, purchased dumpsters for the bairro,

and helped ADASBU register with the government for non-profit status, which qualified it for

CMM contracts. In 2004, the CMM extended one of its first contracts to ADASBU for

USD$1,000/month to continue primary collection in Urbanização (Kruks-Wisner 2006, 44).

81 The 2000 floods exacerbated the already precarious health and sanitation of Maputo, a city with a sewage system that had not been updated since the 1940s, and a population that had grown due to war and subsequent migration to the cities.

82 Joaquim Stretz, AGRESU interview in Maputo at the Feira Ambiental, September 21, 2008.

115

Associação Kithunga originally began as a volunteer neighborhood initiative in September

2006 of 76 members of the Organization of Mozambican Women, the QDWLRQDOZRPHQ¶VPDVV

organization.83 They started collecting trash and raising awareness within the neighborhood about

the risks of cholera and malaria, collected the trash once a week, and swept the roads twice daily.

The Association started working with the Urban District 2 town councilor to involve the CMM.

:LWKWKHWRZQFRXQFLORU¶VVXSSRUWWKH$VVRFLDWLRQQRZKDVD&00FRQWUDFWIRUSULPDU\FROOHFWLRQ

in Urban Districts 2 and 5.84

This section discussed how the expression of voice throughout the city and strong civil society

organizations, with the assistance of donors, positively improved municipal performance of trash

collection in Maputo by extending its reach as municipal contractors. Seed funding from donors

and technical assistance in formalizing their organizations either as non-profits or for-profits was a

critical component in helping these organizations become financially sustainable as city contractors.

Donors helped these organizations lobby the CMM for access to contracts and this successful effort

OHGWRDQH[WHQVLRQRIPXQLFLSDOWUDVKFROOHFWLRQWRLQIRUPDODUHDV,QFRQFOXVLRQ0DSXWR¶VFLYLO

society organizations were strong because they were financially sustainable and were able, in the

instance of ASSOSTI and neighborhood trash organizations, to influence municipal policy.

Merchant Associations had access 10 percent payback for market improvements. ASSOTSI had

access to member dues and neighborhood trash organizations had access to municipal contracts.

83 Xipamanine is a neighborhood that consists of much of the cities commercial activity. The residential areas are informal, cinder-block homes lining winding, narrow dirt roads, tucked behind the commercial areas. The Association was founded in 2006 by a long-time neighborhood resident, vendor, and Xipamanine Market Merchants Association treasurer.

84 Furthermore, as a non-profit organization, women have access to work by paying a membership fee of 25 MTn per month which generates revenue the Association uses towards its activities. They have a calendar in which they do regular cleaning campaigns and have divided the trash clean-up among 3 zones with the neighborhoods where they work. They are looking for funds to purchase a garbage truck and also campaigning for trash drums on the sidewalks for vendors. Although they are working with a national NGO, unlike other microenterprises or community organizations, they did not have a relationship with a donor who advocated for them. They advocated for themselves. Even when local bairro politics conspired to keep the Association from proceeding with its work, the founder directly approached the town councilor for her district and, in this way, won the attention and support of the CMM.

116

All three influenced municipal policy. However, that strength did not in the case of ASSOTSI

extend to better municipal service to informal vendors.

Beira

I expected that civil society organizations would be able to exert pressure over the

municipality to provide vendors and citizens better services in Beira. Specifically, I hypothesized

WKDW%HLUD &0% ¶VFLYLOVRFLHW\RUJDQL]DWLRQVZRXOGEHVWURQJGULYLQJWKHKLJKOHYHORI

performance I observed in the two test areas (see Table 4.2). For both vendor services and trash

collection, I found relatively weak civil society organizations that acted as facilitators of municipal

policy rather than lobbies or self-sufficient service providers. For vendor services, I found that

Merchant Associations facilitated CMB decisions. For trash collection, I found that while

neighborhood organizations provided primary trash collection in some informal neighborhoods,

these groups expanded and contracted based on the availability of donor funding. Both

observations suggest that factors other than civil society organizations ±namely, an increase in own-

source revenue and mayoral leadership ±GURYH%HLUD¶VSHUIRUPDQFHLQWKHDUHDVRIWUDVKFROOHFWLRQ

and market management. Chapters 5 and 6 will explore those factors.

Merchant Associations

8QOLNH$VVRFLDWLRQVLQ0DSXWR%HLUD¶V$VVRFLDWLRQVDUHQRWILQDQFLDOO\VXVWDLQDEOH²

specifically they did not have access to revenue they could use at their discretion. Evidence

suggests that the absence of institutional factors present in Maputo²like regular standard elections,

access to fee revenue and transparent, consistently-applied fee schedules²were reasons why

Merchant Association leadership in the markets I studied does not have a sense of ownership over

market management and is not as strong as MDSXWR¶V0HUFKDQW$VVRFLDWLRQV5HJXODUVWDQGDUG

elections with defined terms would allow Association Presidents to plan changes within the market

and be incentivized to carry them out. Instead, the CMB called elections at different times in

117

different markets, resulting in uneven terms. Second, the fee schedules and market regulations were

outdated and inconsistently applied, so there is little transparency in market management or fee

collection, discouraging vendors from paying because they viewed the fees as unfair or arbitrary.85

The Associations had no control over any portion of the vendor fees, so Presidents had no incentive

to assume ownership over improving market conditions, and fees were irregularly assessed.

However, they were active and partnered with the CMB to implement market changes decided by

the municipality, which may be the reason for improved market infrastructure in the main market.

Interviews revealed that improvements in the markets were associated not with the work of

Merchant AVVRFLDWLRQVEXWUDWKHUZLWK%HLUD¶V0D\RU'DYL]6LPDQJR8QOLNHRWKHU0HUFKDQW

Association leadership in the other cases who held in their positions for years, the Merchant

Association presidents in Beira came into power after the election of Simango, implying a shift in

mayor but also a shift in management favorable to that mayor. Vendors reported a pronounced

improvement in services since 2003, when Simango became mayor. The Central Market

Association adjunct indicated that, since 2003, he noticed a huge improvement in sanitation.86

85 Elections are an example. Although mandates are supposed five years in length, the CMB appears to not consistently apply these rules, which weakens the sanction of the vote since it is infrequently used and seemingly arbitrarily applied. Although the DMF organizes the elections, the mandates are largely decided by the market chief and the individual associations in each market. In practice, the markets have different numbers of members, which also includes the section chiefs, or vendors who represent specific sections by product. Terms of presidents vary. The president (also referred to as chief in Beira) RI*RWR¶V0HUFKDQW$VVRFLDWLRQKDGVHUYHGRQH\HDUDQGZKHQDVNHGKRZORQJKLVWHUPZDVKHODXJKHG DQGVDLG³LWGHSHQGV´7KHPDUNHWFKLHI¶VUHVSRQVLELOLWLHVDUHWRUHVROYHGDLO\SUREOHPVKHOSWKH Merchants Associations keep the inside of the markets clean. The CMB pays for energy, infrastructure and police at formal markets and the bathrooms are the responsibilities of the vendors. Author interview, Town Councilor for Markets and Fairs, Directorate of Market and Fairs office, Beira, March 10, 2009.

86 Before 2003, the sanitation in the market was poor7KHDGMXQFWVWDWHG³ZHKDGWRFORVHWKH bathrooms in the market becaXVHLWZDVGLVJXVWLQJ´$XWKRULQWHUYLHZ$GMXQFW3UHVLGHQWRIWKH Merchants Association, Central Market, Beira, March 11, 2009.

118

Merchant requests to the CMB have been fulfilled and municipal police no longer stole from the

vendors.87

CMB-led Market Shift

While the Merchant Associations I studied in Beira were not financially sustainable, these

Associations facilitated CMB-led market infrastructure improvements as a result of strong mayoral

leadership as the following discussion indicates, making these civil society organizations active

(see Table 4.2). Evidence that market improvements, specifically new market infrastructure, was

more the result of mayoral leadership rather than Merchant Associations surfaced when examining

market renovations in Beira. Plans to improve the Maquinino market started under the previous

mayor. However, Maquinino and Goto Merchant Association leadership credited the current mayor

with the improvements since it was under his mandate that the majority of the market

improvements occurred. Mayor Simango took the lead for making changes in the market without

notable lobbying pressure from vendors. First, the political context of being an opposition city

FUHDWHGDSDOSDEOHIHHOLQJDPRQJYHQGRUVWKDWWKH\KDYHWREDFNWKHLU³KRPHWRZQKHUR´EHFDXVHRI

IUHTXHQWPHGLDDVVDXOWVDQG³VSLHVIURP0DSXWR´7KLVPDGHDFKDOOHQJHWR6LPDQJRHYHQLQWKH

form of a legitimate complaint from vendors, highly unlikely. Second, the population generally

VXSSRUWHG6LPDQJREHFDXVHKHOLWHUDOO\³FOHDQHG-XS´WKHFLW\E\LPSURYLQJWUDVKFROOHFWLRQZLWKLQ

the first six months in office. Third, in appointing his cousin to be the DMF (Directorate of Market

and Fairs) town councilor, Simango took control of the markets and increased fee collections

through a trusted proxy. Fourth, the CMB undertook several infrastructure improvement projects in

Maquinino, thereby providing visible service to the vendor population. Lastly, the nature of the

frequent moves of vendors between Goto and Maquinino as a result of infrastructure improvements

blurred the line between formal and informal vendors in the two largest markets, thereby

87 Specifically, the adjunct indicated that cleaning and water proposals have been accepted. Ten years ago, the municipal police used to steal from the vendors but the adjunct related that this no longer occurs and the CMB-appointed a night guard. When I asked if he feels the presence of the CMB in the PDUNHWKHLQGLFDWHGWKDWWKHVHUYLFH³GRHVQRWIDOWHU´$XWKRULQWHUYLHZ$GMunct President of the Merchants Association, Central Market, Beira, March 11, 2009.

119

discouraging the creation of a particular lobby. These factors combined to produce weak Merchant

Associations because there was no incentive to protest against a popular mayor when visible

improvements to the markets were being made.

The vendor relocations and massive market improvements made to the formal market of

Maquinino illustrate this dynamic. By 1994, vendors had outgrown Maquinino and started moving

to the informal market of Goto to sell fruits and vegetables, resulting in overcrowding and poor

sanitation.88 When Mayor Simango assumed office in 2004, he directed his cousin to add the

necessary infrastructure and divide the markets according to food stuffs (Maquinino) and non-

edible items (Goto). These changes required that sections of Maquinino market be temporarily

emptied. The former town councilor orchestrated successful temporary relocations of vendors from

Maquinino to Goto in order to conduct extensive renovations. As the DMF town councilor, his

office started by informing vendors through notices in the market a month in advance of the plans

to move and why such a move was necessary. When the day came, DMF staff in both markets

facilitated the move. There was little vendor resistance, which interviews with the Merchant

$VVRFLDWLRQVLQGLFDWHGZDVDWWULEXWDEOHWRWKHYHQGRUV¶EHOLHIWKDW0D\RU6LPDQJRZDVDFWLQJLQ

their best interests.89 Since 2004, the CMB has added a section of approximately 200 additional

stalls, two roofs, water pumps and some drainage lines. As the sections were completed, vendors

moved back in waves from Goto to their new stalls in Maquinino.

This episode demonstrates that the CMB took a lead role in vendor relocation and

infrastructure development at Maquinino. The Maquinino Merchant Associations played only a

88 Both Goto and Maquinino have existed in Beira since the 1980s, with parts of Maquinino dating back to colonial times.

89 He successfully challenged the incumbent FRELIMO mayor and was the first opposition mayor to ZLQ%HLUD7KHSHUFHSWLRQZDVWKDW6LPDQJR¶VWRRNDULVNLQFKDOOHQJLQJWKHGRPLQDQWSDUW\DQGKHGLG VR³IRUWKHSHRSOH´

120

nominal role in this process.90 :KLOHLPSURYHPHQWVLQ0DTXLQLQRFRPPHQFHGSULRUWR6LPDQJR¶V

tenure, Association leadership attributed the success iQ0DTXLQLQR¶VUHQRYDWLRQWR6LPDQJR¶V

³VHULRXVOHDGHUVKLS´:KLOH0HUFKDQW$VVRFLDWLRQVLQ%HLUDZHUHQRWVWURQJEHFDXVHWKH\ODFNHG

both financial sustainability and influence to change municipal policy, they were active facilitators

and partners with the CMB in providing new market infrastructure to vendors.

Beira Trash Collection

3ULRUWR%HLUDZDVNQRZQDVWKHXQKHDOWKLHVWFLW\LQ0R]DPELTXH ³SRUFHQWR´

2005).91 So how did the CMB convert itself to a national model? Strong leadership from Mayor

Simango combined with donor support (Folha Informativa 2002; Beira Já Tem Meios 2006). 92 As

90 Furthermore, in discussions with the current Merchant Association leadership in both markets indicate that the vendors were satisfied with the CMB improvements.

91 In 2005, its HIV/AIDs infection rate was 40 percent. Sofala province, of which Beira is the capital and largest city, had an average malarial fatality rate of 12 percent, 3 percentage points above the national average, and cholera fatality of 19 percent, 10 percentage points above the national average from 2002 to 2007. The fact that most of Beira lies below sea level in a tropical climate and the water sanitation system dates back to colonial times also makes it prone to flooding.

92 The Municipal Directorate of Urban and Environmental Management is charged with providing sanitation and funerary services. It is composed of two departments and three services which include the Department of Urban Services, which has over 690 functionaries, where the Department of Solid Waste Management sits (DSWM). Unlike Maputo, the CMB conducts almost all the trash collection directly, without the use of private contractors. The CMB offers residential, hospital and commercial trash collection services and charges an array of prices for each service. Mayor Simango had taken trash collection very seriously. In one example, as late as 2006, trash was still accumulating in piles on the streets, which caused a fire. The CMB purchased a $100,000 USD dump truck and $130,000 USD IODWEHGFRQWDLQHUWUXFNWRLQFUHDVHWKHFLW\¶VFDSDFLW\WRUHPRYHJDUEDJH)XUWKHUPRUHWKH&0% purchased a $40,000 USD tow truck to work exclusively in the shantytowns. Mayor Simango HPSKDVL]HGWKDWIXQGVZHUHIURP%HLUD¶VRZQWD[FROOHFWLRQVDQGWKDWWKHEXLOGLQJRIDQDGGLWLRQDO market was being postponed in order to accommodate this emergency response.

Another factor that PD\KDYHFRQWULEXWHGWR0D\RU6LPDQJR¶VDELlity to effectively and quickly LPSURYHWKH&0%¶VSHUIRUPDQFHRIWUDVKFROOHFWLRQZDVDIRUWXLWRXVFKDQJHLQWKHSULFHRIJDUEDJH containers. During the mandate of Mayor Muchangage, Beira had to purchase containers from Maputo, which where both expensive and were accompanied by high transportation costs. During Mayor 6LPDQJR¶VPDQGDWHDIDFWRU\LQQHDUE\&KLPRLRUHGXFLQJFRVWVRQERWKIURQWV,QWHUYLHZVZLWKWUDVK workers confirm rapid improvements in trash collection. According to one of the trash teams, within WKUHHPRQWKVRI0D\RU6LPDQJR¶VHOHFWLRQWUDVKFROOHFWLRQLPSURYHGEHFDXVHWKHSURFHVVEHFDPHPRUH regimented and organized with the adherence to schedules and planned routes. The CMB has approximately 364 trash collectors who work in teams in three-month shifts. They work three shifts during the day, which provides approximately 20 hours of coverage. However, the improvement, according to these workers, in city services began in 1999 with the introduction of great local autonomy 121

ZLWK%HLUD¶V0HUFKDQW$VVRFLDWLRQVQHLJKERUKRRGRUJDQL]DWLRQVZHUHQRWVWURQJHQRXJKKHUHWR

play a regular role in the improvement of trash collection because they were not financially

sustainable and they did not influence municipal policy in such a way as to consistently improve

trash collection. However, these civil society organizations were active because they did sweep

streets and clean drains in informal neighborhoods when donor funds were available. Like Maputo,

donors worked with neighborhoods organizations to improve trash collection. Unlike Maputo, the

partnership between neighborhood organizations and donors in Beira did not build their

organizational capacity to become contractors and, thereby, become financially sustainable

extensions of the municipality. As will be discussed, it was mayoral leadership and own-source

revenue increases that improved trash collection in Beira

All focus groups concurUHGWKDWWKH&0%¶VWUDVKFROOHFWLRQVHUYLFHKDGJUHDWO\LPSURYHG

almost immediately under Mayor Simango. Palmeiras residents acknowledged that they saw a

larger improvement between the Muchangage and Simango Administrations than they saw since

1999. 93 These participants saw Simango as personally responsible for the change and attributed

UHLQYLJRUDWHGFLW\VHUYLFHVWRKLVRZQSHUVRQDOFRPPLWPHQW³3HRSOHDUHPRWLYDWHGQRZEHFDXVHRI

as a result of decentralization. The key difference the trash collectors identified between 1999 and 2003 was that they began to see better trash vehicles and the salary has been regularly paid on time.

93 Located in Zone 1, Paleiras I borders Chipangara (Inhamodima) to the east and Ponta Gêa to the west. Unlike Inhamodima or Munhava, it is extremely difficult to find volunteers for this focus group during the week because many of the residents are working full-time and are not available. The Secretary of the bairro, is new and explains that the neighborhood is not particularly galvanized because residents are high-LQFRPHDQG³FDQ¶WEHERWKHUHG´)XUWKHUPRUHWKHUHVLGHQWVRIWKLVQHLJKERUKRRGDUHNQRZQDV FRELIMO party supporters, especially since the Pedagogic University dormitory for professors is located in this neighborhood. However, we are able to gather seven individuals to participate (three women and four men). Author interview, Secretary of Palmeiras I, Beira, March 17, 2009.

Although participants acknowleGJHWKDWLWLVWKH&0%¶VUHVSRQVLELOLW\WRXUEDQL]HDQGNHHSWKH city clean, three participants immediately acknowledge that the CMB has improved its trash collection service. One participant indicated that the means with which the CMB has to do trash collection, namely trash trucks and containers, has visibly improved. Six to seven years ago, open-air defecation was a regular occurrence. One resident pointed to a street near the meeting site to share that it was completely inpassable because of the mountain of trash. Participants indicated that they noticed a slight LPSURYHPHQWGXHWRGHFHQWUDOL]DWLRQLQ2QHSDUWLFLSDQWVWDWHG³WKHFLW\LVZRUNLQJODWHO\,WLV HYLGHQW´³7RGD\\RXGRQ¶WVHHWUDVK´DQRWKHUSDUWLFLSDQWVKDUHG3DUWLFLSDQW3DOPeiras I focus group, Beira, March 17, 2009.

122

Simango. Nobody before wanted to contribute but now, there is politiFDOZLOO´VKDUHGRQe

participant.94 )XUWKHUPRUHSDUWLFLSDQWVWUXVWKLPZLWKSXEOLFIXQGV³:HGRQ¶WZDQWSHRSOHZKR

take the money and put it in their pockets. (Simango) takes the money and does something with

LW´95 Most Ponta Gêa96 participants concluded that trash collection was an area where they had seen

vast improvement, attributing this improvement to Mayor Simango. Despite the dissatisfaction

with paying the fees, participants acknowledged that the situation had changed for the better.97

While focXVJURXSVLQGLFDWHGWKDWWKH&0%¶VSHUIRUPDQFHRIWUDVKFROOHFWLRQKDGLPSURYHG

local civil society organizations, specifically neighborhood associations, were not the main factors

GULYLQJWKLVSHUIRUPDQFHEXWUDWKHUGRQRUVDQGWKHPD\RU¶VGLUHFWLQWHUYHntion. Instead of

0DSXWR¶VVLWXDWLRQZKHUHFRQWUDFWXDOUHODWLRQVKLSVFUHDWHGDPDUNHWIRUPLFURHQWHUSULVHVWKH&0%

provided no incentive to neighborhood organizations and no avenue for sustainable income. As

94 Participant, Palmeiras I focus group, Beira, March 17, 2009.

95 Participant, Palmeiras I focus group, Beira, March 17, 2009.

96 Located in Zone 1, Ponta Gêa is comprised of five sub-bairros or units (A-E), totaling 18,382 residents, with Unit E, where Goto market is located, being the most populous. The neighborhood is bordered by Samora Machel Avenue, the beach, Armando Tivane Avenue on the north, and July 24th Avenue. Each unit has a unit secretary who then oversees the block chiefs who oversee the chiefs of ³KRXVHV´ZKLFKLVWKHVPDOOHVWXQLWRIRYHUVLJKW&XULRXVO\LWLVDOVRWKHQHLJKERUKRRGZLWKERWK the largest informal market (Goto) and the largest informal settlement, the Grand Hotel, in what was once a high-end hotel with approximately 3,000 residents, its own school, mosque, church, and internal ³SROLFH´IRUFH$XWKRULQWHUYLHZ6HFUHWDU\RI3RQWD*rD%HLUD0DUFK

The group also identifies cleaning ditches and drains, preventing sand dune erosion, improving the health of residents, urbanization, signage, public bathhouses, roads and schools as CMB responsibilities. One participant astutely recognized that there are areas, such as coastal erosion and schools, where the central government and local government overlap in responsibilities, with can lead to confusion, a theme identified in the Palmeiras focus group. Consistent with scorecard findings, participants indicated that there needs to be a better exchange among residents and that they cannot completely rely on the CMB to do all the work. Four of the ten respondents indicated that, as a second option, if going to the CMB did not yield results, they would organize a group in their own neighborhood to pick-up the trash.

97 Two participants indicated that if the service were better, they would pay more. Some indicated they did not know how the fees were used. Another participant indicated that funds go the central government. Like Inhamodima residents, Munhava participants indicated that they do not see the service in the bairro2QHSDUWLFLSDQWLVIUXVWUDWHG³$OPRVWHYHU\RQHKDVHOHFWULFLW\DQGLWLVZHOO organized and we pay the trash fee. But we have to dig a hole to bury our trash because the car does not HQWHULQWRWKHFRPPXQLW\7KH&0%VKRXOGFRPH:HSD\IHHVDQGWKH\GRQ¶WSLFNXSWKHWUDVK´ Participant, Munhava Central Focus Group, Beira, March 14, 2009

123

such, these organizations rely largely on donors to piece together the equipment needed and service

to informal neighborhoods remains irregular. Despite these conditions, focus group participants

VLJQDOHGFRQILGHQFHLQ0D\RU6LPDQJR¶VDELOLW\WRFOHDQXSWKHFLW\EHFDXVHRILPSURYHPHQWVWKH\

noticed outside their neighborhoods. 98 While donors were working with the CMB to improve

trash collection, unlike in Maputo, these efforts were not always coordinated with the CMB and

neighborhood leadership in a way to make these efforts a sustainable extension of the municipality,

thus preventing improvement in municipal capacity and performance. Evidence from two

neighborhoods illustrated how the lack of incentives and capacity for residents to organize, external

DFWRUV¶PLVFRPPXQLFDWLRQZLWKUHVLGHQWVDQGDlack of a contractual connection to the CMB

created and facilitated unsustainable and fragile associations. 99

Weak Neighborhood Associations: Influence of External Actors

As previously stated, neighborhood associations in Beira were weak because they were not

financially sustainable and did not have influence over municipal policy to improve trash

collection. They were active, however, because external actors²specifically donors and the

98 We asked residents to rank the top three actions they would undertake to improve city services. The majority of the respondents indicated great confidence in voting. Eight of the eleven respondents ranked first that they would participate in the electoral process, consistent with the scorecard finding that it is worth participating in the electoral process. Half of the respondents ranked second that they would organize a group to press the CMB to improve service, again revealing that residents expect the actions to improve their trash collection to be undertaken by the CMB primarily. Despite the responses that demonstrated that focus group participants see the CMB as the primary trash collection service provider, in the comments section, four of the seven comments requested that a group of ladies be informally organized to collect the trash within the neighborhood. Interestingly, ranked second and third E\VRPHUHVLGHQWVZHUHWKHIROORZLQJWZRRSWLRQV³RUJDQL]HDJURXSZLWKLQP\QHLJKERUKRRGWRSLFN XSWKHWUDVK´DQG³SD\VRPHRQHLQIRUPDOO\WRSLFNXSWKHWUDVK´7KLVILQGLQJLVFRQVLVWHQWZLWKWKH bairro¶VH[SHULHQFHZLWKJURXSV,QWKHVFRUHFDUGH[HUFLVHPRVWUHVSRQGHQWVDJUHHGWKDWSDUWLFLSDWLQJLQ DPXQLFLSDOPHHWLQJLVDQHIIHFWLYHPDQQHURIFDOOLQJWKH0D\RU¶VDWWHQWLRQWRDSUREOHP6Rdespite the complaints that trash collection within the bairro should be greatly improved, overall, respondents have DJUHDWGHDORIFRQILGHQFHLQWKH&0%¶VLQVRIDUDVWKH0D\RULVFRQFHUQHG$QRWKHUJURXSH[HUFLVH reiterated these findings. Consistent with this first choice, residents selected that the CM has a good process for consulting with the residents and that it is worth participating in the electoral process.

99 Some residents were skeptical, not of the CMB, but of the bairro OHDGHUVKLS¶VLQDELOLWy to deliver and organize the community into projects that are sustainable. One participant indicated that, if an individual starts picking up the trash, residents assume that person is in charge and give the garbage to them. Raising proposals to the CMB is complicated, one participant states, because they feel so isolated.

124

Catholic University of Beira (CUB) ²sporadically provided financing and supplies. The support I

observed did not build the kind of sustainable relationship between the municipality and

organization present in Maputo to be able to extend municipal service to improve trash collection.

Inhamodima and Munhava shared a common factor: external actors greatly influenced the

creation and sustainability of community organizations in the area of trash collection. Although

0XQKDYD¶VH[SHULHQFHLVPRUHSRVLWLYHWKDQ,QKDPRGLPD¶VXQOLNHZLWK0DSXWR¶VPLFURHQWHUSULVHV

community organizations in both neighborhoods were unable to maintain a sustainable level of

operations because of a lack of coordination between the municipality and the CUB. In Munhava,

the donor was working closely with the CMB as a part of a major overhaul of the drainage system,

which required high levels of coordination between both the municipality and the donor. This

condition was absent in Inhamodima. The miscommunication in Inhamodima between the Catholic

University staff and neighborhood leadership persisted and, ultimately, resulted in the end of this

project, with no discernable intervention from the CMB.

The CUB began a project in 2005 in two subsections (units A and B) of Inhamodima, with 30

volunteers from the neighborhood. The purpose of the project was to organize the community to

volunteer to carry out primary trash collection in the neighborhood. Over the two years the project

ran, volunteers swelled from 30 to 200 individuals.100 Volunteers expected to be remunerated, an

expectation stoked by the bairro Secretary who implied that the CUB would bring donor funds to

eventually pay residents.101 Although the CUB had paid for materials, remuneration was not

100 The first 30 volunteers were from the community benefiting from the clean-up. Individuals in the surrounding area who were not benefiting from the collection but rather looking for work caused the increase in volunteers.

101 In my own interview with the Secretary, he mentioned various donor interests in Inhamodima at least four separate times. In one of the final discussion groups the CUB conducted, the majority of the discussion centered around trash collection but curiously, when the discussion closed no one was interested in volunteering to follow-XS3DUWLFLSDQWVIHOWLWZDVHLWKHUWKH&8%¶VRUWKH&0%¶V UHVSRQVLELOLW\,QIDFWDFXOWXUHH[LVWVDPRQJWKHUHVLGHQWVRI³(SD7HP SURMHFWRQREDLUUR´RU³+H\ WKHUH¶VDSURMHFWLQWKHQHLJKERUKRRG´ZKLFKLVDZD\WRJLQXSLQWHUHVWLQJHWWLQJUHVLGHQWVWRZRUNLQ the hopes of being compensated. Interestingly, one participant recalled that, during colonialism, trash 125

included.102 By March, 2007, CUB organizers, and community residents held a cantankerous

meeting in which volunteers demanded that they be paid. They requested USD$2,000 per volunteer

per year, an amount the CUB had neither promised nor budgeted.103 The meeting ended poorly,

with residents feeling insulted and the project defunct. 104

:KLOH,QKDPRGLPD¶VH[perience with donors resulted in misunderstandings that terminated the

SURMHFW0XQKDYD¶VH[SHULHQFHGLGQRW,QWHUYLHZVVXJJHVWWKDW$IULFD¶VLQYHVWPHQWLQ

0XQKDYD¶VGUDLQDJHV\VWHPUHTXLUHGFRRSHUDWLRQZLWKWKH&0%QHLJKERUKRRGOHDGHUVKLSDQG

neighborhood organizations. Africa 70, along with the UNDP, undertook a major drainage project

requiring a multiyear, multistage effort with the CMB. A key component of this project was

coordinating with neighborhood organizations. As a result, Africa 70 conducted a census of

neighborhood organizations and began identifying the ones it could fund due to its funding

availability. This NGO was co-located with the neighborhood leadership and, when

misunderstandings arose, Africa 70 staff worked to correct them immediately. This coordination

between the CMB, neighborhood leadership and neighborhood organizations, while not

LPPHGLDWHO\JHQHUDWLQJVXVWDLQDEOHRUJDQL]DWLRQVWKDWFRXOGHLWKHUH[WHQGWKHPXQLFLSDOLW\¶VUHDFK

or fill in the gaps over time, did provide some short-term primary trash collection and drainage

clean-up.

pick-up did indeed occur in the bairro . It appeared as if a tractor could indeed fit down the main unpaved road through the center of this bairro. Author interview, Chipangara Bairro Secretary, Beira, March 12, 2009.

102 One of the volunteers actually stole new materials CUB provided and sold them. When CUB staff finally tracked him down, he promised to pay but then left the neighborhood.

103 One student present at the meeting opened up his wallet and threw money at the CUB organizers, trying to shame them into paying these people who he claimed were suffering and were mislead.

104 Evidence of how much residents rely on external assistance surfaced in a later CUB-sponsored discussion group. CUB-sponsored discussion group, Inhamodima, Unit B, Beira, March 21, 2008.

126

Africa 70 identified over 20 community associations in Munhava.105 These associations raised

awareness in the neighborhood regarding various aspects of sanitation and health. VIDEC conducts

primary trash collection, clears drains and raises awareness about the need for proper waste

disposal.106 External actors played a pivotal role in establishing and supporting VIDEC. ESSOR, a

)UHQFKRUJDQL]DWLRQZRUNLQJRQLPSURYLQJ0XQKDYD¶VVDQLWDWLRQVXJJHVWHG that the residents form

a group that focuses on drainage and sanitation, resulting in VIDEC.107 However, it lacked

sustainable support to continue activities. Then, in 2007, VIDEC began working with CUAMM, an

Italian NGO, to build ecological latrines and by 2008, VIDEC began working with Africa 70 to

identify the families that could afford to partially subsidize the latrines. Today, VIDEC contracts

residents to act as door-to-door activists and provides a small subsidy for these activities.108 Like

0DSXWR¶V trash collectors, the VIDEC president observed that people were very suspicious towards

105 There is an informal group of young people who clean the drains called the Youth League that works in the 12th sub-bairro. Another group, Fight for Life (Luta Pela Vida) is another community organization that takes care of orphan children in Munhava Matope, one of the poorest sub-bairros in Munhava, located adjacent to the dump. Two organizations, Light in the Community (Luz Na Comunidade) and Child Orphans Auxiliary (Auxilio das Crianças Orfeas) are religious organizations that work within Munhava to assist individuals with HIV/AIDS and orphans. These organizations also extend their work to promote the use of latrines, cleaning of water to reduce transmission of cholera. %RWKSDVWRUV,LQWHUYLHZHGLQGLFDWHGWKDWWKHLUELJJHVWFKDOOHQJHLVFKDQJLQJSHRSOH¶V behavior in light of a lack of infrastructure and resources.

106 VIDEC has about 24 members, most of whom are drawn from Munhava Central, who work on cleaning the drains and raising awareness about the connection between sanitation and sicknesses like cholera, typhoid and malaria. Both city staff and VIDEC staff emphasized the legitimacy of the organization-that its existence and bylaws were published in the Mozambican National Federal Register or Boletím da República DQGWKHVWUXFWXUH9,'(&¶VVWUXFWXre²which contains an Assembly, a Fiscal Committee and Executive Council²is a structure similar to many other Mozambican NGOs. I was struck by how much of a focus is placed on formalizing organizations and how the organizations then founder. Interestingly, almost all Mozambican community organizations with whom I spoke said they collected dues. Members are charged on average 10 MTn/month to participate and can be charged an initiation fee. However, these fees do not cover the costs of their activities.

107 Open-air defecation was a major issue given that, for example, there were only two public bathrooms in two Munhava sub-bairros that served between 8,000 to 10,000 residents.

108 7KH9,'(&SUHVLGHQWLQGLFDWHGWKDWVLQFHKHKDVVHHQ³WKLQJVVWDUWHG WRPRYH´³7KH&0%LV ZRUNLQJDQGVRLVWKHFRPPXQLW\´KHVKDUHG,QWHUHVWLQJO\XQOLNH,QKDPRGLPDZKHUHH[WHUQDODFWRUV ZHUHSHUFHLYHGWRKDYHGDPDJHGWKHFRPPXQLW\¶VGHVLUHDQGDELOLW\WRRUJDQL]HWKH9,'(&SUHVLGHQW observed that donors in Munhava KDYH³KHOSHGSV\FKRORJLFDOO\FKDQJHSHRSOH¶VZD\RIWKLQNLQJ´E\ identifying problems within the community that could galvanize residents. Author interview, VIDEC President of the Assembly, Munhava Central bairro office, Beira, March 26, 2009.

127

9,'(&¶VUHSUHVHQWDWLYHVZKHQWKH\GLGWKHLUGRRU-to-door campaigns.109 Six years later, VIDEC

has been able to overcome these perceptions because they live in the community and are known by

the bairro leadership. 110 However, they still expand and contract based on available donor

funding.

A major difference between Maputo and Beira was that in Maputo, donors both provided

VXSSOLHVKHOSHGEXLOGWKHFLYLOVRFLHW\RUJDQL]DWLRQV¶FDSDFLW\WRDFWDVFRQWUDFWRUVIRUWKHFLW\DQG

helped lobby the CMM to provide these organizations contracts. In Beira, this was not the case.

Funding from external actors was sporadically available and did not create financially sustainable

RUJDQL]DWLRQVLQWKHDUHDRIWUDVKFROOHFWLRQ,Q,QKDPRGLPDWKH&8%¶VSURMHFWHQGHGLQDPDVVLYH

breakdown in a large bairro initiative because communication was not clear about the scope of the

project and its purpose. In Munhava, although community organizations existed and worked within

the community, their activities were largely influenced by sporadically available donor funding. In

both instances, municipal performance of trash collection did not improve but trash collection

109 VIDEC was conducting a survey of those who had electricity and water and one of the resident who was also a member of the military pulled a pistol on them, saying that they would kill them for asking these questions.

110 Another organization in Munhava Central is the Trash Committee, which is comprised of five residents who voluntarily keep the terciary drains clean of rubbish and debris so that the water does not stagnate. It began in 2005 as a result of CUAMM support primarily to educate the population about the spread of cholera but ceased work in 2007 due to lack of gear. They started work again with Africa 70 in March 2009, with gloves, masks, shovels and hand carts to do the work. Like VIDEC, the Trash Committee faces resistance from residents. Lastly, the trend evident in Beira, Matola, Dondo and, to a lesser extent, in Maputo, is that community organizations tend to crop up when there is donor funding and become dormant until external funding is available.

Despite donations of equipment, the Committee is still missing gear, members are not subsidized and equipment requires maintenance. Although these types of organizations are necessary to educate the public on good hygiene, the conditions in Munhava are infamously renowned for lack of sanitation, open-air defecation, lack of public bathhouse, flooded sewer drains. There are not enough public bathrooms and so individuals build impromptu outhouses, which, if they are lucky, they can build near one of the sewage lines so it can drain away from the house. Unfortunately, as the drains are RSHQDQG%HLUDLVDVZDPSWKHZDWHUOHYHOULVHVGXULQJUDLQ\VHDVRQIORRGLQJVHZDJHLQWRSHRSOH¶V homes. Although some of these acts are behavioral, alternatives for the proper disposal of waste and lack of proper sewage infrastructure makes changing behavior difficult. Furthermore, these organizations are primarily active in Munhava Central, the area, which appears to have the highest donor intervention. Other parts of Munhava, such as Munhava Matope near the trash dump, are in far worse shape but do not have the same donor or civil society presence.

128

within the neighborhood temporarily did due to funding from external actors. Unlike the case of

0DSXWR%HLUD¶V,QKDPRGLPDDQG0XQKDYDQHLJKERUKRRGRUJDQL]DWLRQVGLGQRWJHWWHFKQLFal

DVVLVWDQFHWRKHOSWKHPWRDFWDVFRQWUDFWRUVH[WHQGLQJWKH&0%¶VUHDFKDQGLPSURYLQJ

performance.

Dondo

,K\SRWKHVL]HGWKDW'RQGR¶V &0' VWURQJFLYLOVRFLHW\RUJDQL]DWLRQVZRXOGKDYHD

SURPLQHQWUROHLQ'RQGR¶VSHUIRUPDQFHRIWUDVKFROOHFWLRQDnd market management (see Table 4.2).

As I previously indicated, I expected civil society organizations to exert pressure over the

municipality to compel them to provide vendors and citizens better services (see Table 4.1). In

Dondo, the main pressure from residents came not through civil society organizations but rather a

donor-funded participatory budget process the CMD instituted with the help of the Austrian

Development Cooperation. Like Beira, I found that Merchant Associations were not strong

because they were not financially sustainable. However, in the area of market management,

Merchant Associations were active because they facilitated municipal policies and practices.

Specifically, in the larger markets, they actually performed a key municipal function²fee

collection. Dondo was the only municipal case that had decentralized fee collection in its municipal

markets, a move that has shown some promising results.111 Although vendors are not permitted to

keep the collections, they were satisfied with CMD services and their involvement in fee collection

VHHPHGWRSURYLGHDNLQGRI³EX\-LQ´HIIHFWVLPLODUWRWKHSHUFHQWSD\-back in Maputo. In the

area of trash collection, I found that, while civil society organizations were not strong because they

111 The CMD designates markets as municipal (formal) markets, where fee collection is conducted by the Merchants Association, and community (informal) markets, where the CMD conducts the collection. There are 6 municipal markets and four community markets. The 6 municipal markets are Mafarinha, Central, Manual Cambezo, Kumsitu, Kanhangula, and Nhamainga. The 4 community markets are Samora Machel, Pequeno Central, Campo Estado and Mesquita.

129

were neither financially sustainable nor did they influence municipal policy, they did act as

interlocutors between the CMD and the vendors.

Merchant Associations

$VVWDWHGDERYH'RQGR¶V0HUFKDQW$VVRFLDWLRQVZHUHQRWILQDQFLDOO\VXVWDLQDEOH+RZHYHU

they were active because the Merchant Associations in the largest markets were delegated major

municipal function²collection of vendor fees. For these markets, the CMD was deemed

responsive in responding to vendor requests. The CMD was responsible for the formal market

infrastructure in Mafarinha, Central Market, and Manual Cambezo. All three markets had

bathrooms, water taps, and public illumination. Presidents knew who their vendors were, the

amount of collections, and had successfully organized to lobby the CMD for better services. The

Cambezo Association president indicated that the CMD has complied with most requests they have

put forth, such as a water fountain, a bathroom, and cleaning crew.112 When a hurricane blew

chapas onto the roof of Central Market, vendors wrote a letter to the CMD and the CMD removed

the debris and rebuilt the roof. 113 In Mafarinha market, the CMD removed a tree that fell on the

URRIDQGWKHPDUNHW¶VUHTXHVWIRUUHSDLUVKDGEHHQSURJUDPPHGLQWRWKHPXQLFLSDOLW\¶VEXGJHW

Decentralized Markets and Fee Collection

2QHNH\GLIIHUHQFHEHWZHHQWKH&0'¶V0HUFKDQW$VVRFLDWLRQVDQGWKRVHRIRWKHU

municipalities was that Dondo had decentralized fee collection in the major markets. Specifically,

the CMD delegated fee collection--a municipal function vested to municipalities per

decentralization--to the Merchant Associations of the largest markets. Unlike all other cases,

112 0DQXHO&DPEH]R0DUNHWZDVUHQRYDWHGXQGHU&DPEH]R¶VPDQGDWHDQGVRZDVUH-named after him by the vendors. The Association president has access to Mayor Cambezo directly. Central Market is the only other market in which the Association president also has access to Mayor Cambezo. Combined, the two markets generate half of market fee revenue.

113 However, even with CMD support, markets have problems. Both Cambezo and Central Market are centrally located in the populous neighborhoods making stealing a regular occurrence. Central Market has had serious problems with theft and merchants have hired 20 private guards to augment the one police officer who was assigned to guard the market.

130

'RQGR¶V0D\RU&DPEH]RLQLWLDWHGGHFHQWUDOL]HGIHHFROOHFWLRQLQZKLFKYHQGRUVLQVRPHRIWKH

markets collect fees RQWKHPXQLFLSDOLW\¶VEHKDOIHYLGHQFHRIKRZDFWLYHO\HQJDJHGZLWKWKH&0'

the Merchant Associations were.114 Fees are collected twice a day by two members, usually the

SUHVLGHQWDQGDGMXQFWDQGWKHQGHSRVLWHGLQWKHPRUQLQJLQWKHWRZQFRXQFLORU¶VRIILFe. While the

CMD did not have a formal revenue-sharing agreement with these Associations as those in Maputo

did, these Associations had cooperative relationships with the CMD such that they were vested with

the municipal responsibility of collecting fees.

Two issues with this practice surfaced. First, how (and how regularly) collections were audited

remains unclear. Secondly, Association leadership was exempt from paying the fee. This practice

could discourage vendors from paying and also Association leadership from pursuing payment.

However, despite these concerns, estimates of fee collections from the two largest municipal

markets that were decentralized (Central Market and Cambezo) indicated that collections at these

markets were two to three times the average monthly fee collection across all markets.115

Furthermore, comparing actual collections to projected average collections, Cambezo estimates are

almost identical, indicating that the vendors are collecting fees at or near capacity (Mapa de

Controlo 2006).116 These estimates demonstrated that at least in the two largest markets were the

114 Currently, CMD markets are overseen by 2 market chiefs and fees are collected by one of 19 CMD fee collectors. Fee collection and initial dispute resolution is handled by the Merchants Association. Until 2007, all markets were community (or informal) markets. One Merchant Association president has seen many changes in the markets, which include an increase in vendors, diversity and number of products sold, and new stalls. Dondo residents go to Beira much less now than they used to ten years ago because there are more products for sale. Author interview, Local Economic Development staff, December 11, 2008.

115 Local Economic Development estimates for 2008 market fee collections are $48,844 USD, making the market annual average $4,884 USD. Assuming an average of 5 MTn in daily fees and full collection, Central Market generates $13,584 USD/year and Cambezo generates $9,056 USD/year. $XWKRU¶VFDOFXODWLRQVEDVHGRQ&0'GDWD

116 I compared daily estimates based on number of vendors and average fee intake (5 MTn per vendor) DQGGDLO\HVWLPDWHVEDVHGRQRQHZHHN¶VZRUWKRIDFWXDOFROOHFWLRQVDQGLQERWKLQVWDQFHVWKHGDLO\ collection average was $37 USD. Central Market estimates were off by almost $20 USD. However, I attribute this to two factors. First, the number provided me of vendors in Central Market was actually the number of stalls. Second, given that vendors tend to occupy the stall without actually being present, vendors cannot collect the daily fee from stalls in which vendors are absent. 131

function of fee collection had been decentralized, the Merchant Associations were incentivized

enough to compel (and turn in) collections. In conclusion, while Dondo¶V0HUFKDQW$VVRFLDWLRQV

were not financially sustainable, they were entrusted by the CMD to perform an important

municipal function. The partnership established as a result of Merchant Associations collecting fees

allowed the Merchant Associations to work with the CMD to address infrastructure requirements

such as the replacement of market roofs in Central and Mafarinha markets.

Dondo Trash Collection Civil society organizations in the area of trash collection were weak in Dondo because they

were not financially sustainable and did not influence municipal policy. 117 However, of all the

cases, focus groups expressed the highest degree of confidence in the &0'¶VSHUIRUPDQFHQRWMXVW

LQWUDVKFROOHFWLRQEXWRWKHUDUHDV,QWHUYLHZHHVFUHGLWWKHVHLPSURYHPHQWVWR'RQGR¶V

participatory budget process. Initially funded by the Austrian Development Cooperation, Mayor

&DPEH]RLQVWLWXWHG'RQGR¶VSDUWLFLSDWRU\SUocess to stimulate resident feedback on municipal

infrastructure investment. This process encourages residents to vote on neighborhood projects and

track CMD investments. 118 All focus group members immediately identified various CMD

117 Like Beira, DonGR¶VFRPPXQLW\RUJDQL]DWLRQVDUHVRPHZKDWZHDNEHFDXVHWKH\DUHJHQHUDOO\QRW self-sustainable. However, the CMD has arranged a partnership with one community organization that allows the organization to support activities. With ten activists and 250 memEHUV3URMHFW8'$'¶V mission is to raise community awareness regarding hygiene and the dangers of open-air defecation on public health. Although the CMD does not contract with UDAD for these activities, they have allowed them to keep fees they charge for use of public bathrooms. Despite the absence of a CMD contract, the ability to keep bathroom and member fees allowed UDAD to raise almost 6,800 MTn/month ($256 USD) which they use to pay their activists 300 MTn/month ($11.32 USD) with about 400 MTn/month ($15.09 USD) remaining.

118 The informal neighborhood focus group is comprised of 11 men and women from informal neighborhoods, which include Macharote, Mandruze, periphery of Central, Kumsitu, and Nhamainga. The mixed group is a combination of 11 men and women from formal and informal neighborhoods, which include Mafarinha, Nhamainga, Central, Nhamayiabwe, and Macharote. The formal neighborhood focus group was comprised of 7 men from Central bairro. Urban Services and Environmental Management (Ser vi ços Urbanos e Gestão Ambiental or SUGA) is comprised of four units: sustainable development, which controls erosion and general municipal beautification; solid waste services, which conducts trash collection and audits behavior; cementary management, in which the CMD has one Toyota truck used to pick up the deceased for processing at the recently-built morgue; and internal administration, which tracks revenues and expenses and is also in charge of human resources. Author interview, SUGA Town Councilor, Dondo, November 4, 2008.

132

improvements, not only trash collection but also paved roads, running water and new municipal

EXLOGLQJVDWWULEXWLQJ'RQGR¶VVXFFHVVWRDQQXDObairro-level consultations through its participatory

budget process.119 Despite overall satisfaction, all focus group members observed that the informal

neighborhoods still experience problems with inadequate trash collection. Focus group participants

offered that the CMD was doing a good job but that more trash trucks and workers were needed, as

well as advanced technology associated with modern trash collection since the CMD still relied on

rusted drums and a tractor to collect waste. 120

:KLOHLQIRUPDOQHLJKERUKRRGVVWLOOGLGQRWUHFHLYHUHJXODUWUDVKFROOHFWLRQWKH&0'¶V

participatory budget process extends to all neighborhoods of the municipality as a way to both

manage resident expectations by making town councilors accessible to members of the public and

getting citizen proposals on potential infrastructure investments in their neighborhoods. CMD

residents have direct access to high-ranking town councilors.121 Informal neighborhood residents

indicated that they could raise complaints directly to the town councilors. Furthermore, these

residents indicated that the CMD has a history of completing projects, the result of its annual

participatory budgeting process, which instilled confidence in the CMD. In the cases in which

projects were delayed, one participant stated that it is usually the fault of funding delays among

119 Informal neighborhood participants indicated that management had gotten better since 1999, some pointing to the specific work the CMD did on improving the roads. When asked what were the top three municipal responsibilities, participants mentioned the construction of markets, public bathrooms, schools, mortuary, roads, and a school. Participants of the mixed group also pointed to water provision, assistance to orphans and widows, combating erosion and partnerships with donors and other NGOs.

120 According to one participant, trash collection in the informal neighborhood occurs weekly. Like informal neighborhood participants, formal participants indicated other areas in which the CMD improved, like construction of markets and water fountains and the provision of water 24 hours a day instead of just 6 hours a day, Participant, Focus Group, Dondo, November 11, 2008.

121 In fairness, Dondo is a small in terms of population and it is not unusual to see the town councilors DQGPD\RUZDONLQJDURXQGWKHFHPHQWFLW\$OWKRXJKRQHFRXOGDUJXHWKDW'RQGR¶VVL]HPDNHVLWHDVLHU to improve service and have open communication on municipal projects, I argue that consistent leadership and clear communication to residents spread over a fairly large area are primarily responsible for muniFLSDOVHUYLFHLPSURYHPHQWVDVSHUFHQWRI0R]DPELFDQKDYHSRSXODWLRQVOHVVWKDQ'RQGR¶V and Dondo has been a successful model.

133

CMD partners.122 So, while civil society organizations in this area were generally weak, the CMD

instituted a process that allows regular communication between the CMD and residents, which is

evidence of an active civil society.

Weak Community Organizations,

Active Participatory Process: Variation in Sustainable Internal Support

'HVSLWHZHDNRUJDQL]DWLRQVIRFXVJURXSSDUWLFLSDQWVZHUHJHQHUDOO\VDWLVILHGZLWKWKH&0'¶V

SHUIRUPDQFHDQGFUHGLWWKLVWRWKHPXQLFLSDOLW\¶VFRQVXOWDWLRQSURFHVVIn Dondo, the Austrian

Development Cooperation funded its participatory budget process starting in 1999, which continues

WRLQIRUPWKH&0'¶VGHFLVLRQ-making on investments of its infrastructure budget. Each bairro has

a consultative council, which articulates proposals to the CMD for community funding priorities.

The council has a representative to a specific town councilor who in turn is also charged with

representing that neighborhood. They have annual consultations to help formulate the budget

allocations for funds like the FCA.123 Furthermore, the five-year strategic plan is also composed

WKURXJKDFRPPXQLW\FRQVXOWDWLRQSURFHVV,DWWULEXWHGIRFXVJURXSSDUWLFLSDQWV¶VDWLVIDFWLRQZLWK

services as a combination of actual municipal performance and of this unique consultation process,

which helped residents understand the CMD¶VFRQVWUDLQWV

While this process is considered to be a national model, local researchers have concluded that

it is not clear how the input from the neighborhood councils is incorporated into the final resource

DOORFDWLRQVLQ'RQGR¶VPXQLFLSDOEXGJHW 1XYunga, Mosse, Varela 2007, 13-14). As a result, I

classified this civil society in Dondo as active, because the participatory budget process institutes a

122 Although some participants articulated this perspective, group members also offered ways in which service can be improved which included picking up trash at night, instead of during the day. Participant, Informal Focus Group, Dondo, November 11, 2008.

123 Municipal Compensation Fund or Fundo Compensação Autárquica (FCA) is the generally the largest central government transfer dedicated to paying for municipal infrastructure investment.

134

framework of vertical accountability but not strong, because it is unclear the degree to which these

consultations modified municipal decisions and policies towards infrastructure investments.

Matola

I would expect to find either weak civil society organizations or paucity of civil society

organizations in Matola (CMMa) as this was a case where performance in the two test areas was

rated low (see Table 4.2). I found no civil society organizations working in the area of

neighborhood trash collection. In the area of market management, I found weak Merchant

Associations that were inactive. They had neither access to funds nor did they influence municipal

policy in the area of market management. I also found that nepotism in both market management

and trash collection was more pronounced in Matola than in the other three cases, which might

explain the inability for Merchant Associations to even be active partners with the CMMa. Focus

groups indicated that trash collection was selective in their neighborhoods and directed only at

bairro leadership. Municipal staff themselves revealed that market chiefs likely kept funds they

collected and that the CMMa had no process for accurately tracking vendors or collections.

Merchant Associations

Matola Merchant Associations were the weakest of all the cases both because of a lack of

financial sustainability and inability to influence municipal policy in the context of corruption in the

markets that made partnership with the CMMa difficult. Institutionally, Merchant Association

leadership indicated that they did not hold regular elections, were unaware of the number of their

members, and observed few notable improvements in most markets since decentralization.124 The

fee schedule and fee charging practices were irrational and open to abuse. In 1998, the vendors

held a strike to object to imposition of fees. However, unlike in Maputo, the strike was futile and

124 Madruga did undergo renovations under Mayor Tembe but vendors with whom we spoke generally agreed that after he died, the municipality forgot about the market.

135

fees were imposed. The 2006 fee increase created a higher, flat fee in both formal and informal

markets, which was perceived as unfair by informal vendors but the Associations indicated that

they were powerless to change it. Additionally, fee collectors vary in their practice of charging fees

to informal vendors. Unlike Maputo, where fee collectors did not charge informal vendors in order

to avoid justifying their right to occupy that space, Matola fee collectors charged informal vendors

despite the fact that the police came shortly thereafter to clear those vendors out. This caused

vendors to mistrust the CMMa and its agents. One staffer speculated that fee collectors kept these

³H[WUD´IHHVDVDSD\PHQWWRDOORZWKHYHQGRUWRVHOOLQWKDWVSDFH:LWKWKHH[HPSWLRQRI3DWULFH

Lumumba, where there is a cashLHU¶VER[DQGGHVLJQDWHGKRXUVIRUIHHFROOHFWLRQLWZDVXQFOHDU

how fees or fines were systematically collected.

Associations were weak in both formal and informal markets and did not organize to request

better services because of the following two institutional gaps: a lack of regularized elections and

codified fee schedules combined with corruption in the markets. The lack of market infrastructure

LQPDMRUPDUNHWVSURYLGHGHYLGHQFHRIWKH0HUFKDQW$VVRFLDWLRQ¶VLQDELOLW\WRHOLFLWVHUYLFHVIURP

the CMMa, which itself did not have the staff capacity to track its market fees. Malhanpsene

market has no infrastructure or services. Patrice Lumumba, which generated 9 percent of market

revenue, had no running water after seven years. Maquinino market was saddled with a 65,000

MTn debt to Aguas de Moçambique, which the CMMa stopped paying. Furthermore, the roof was

in need of repair. The Association reported that they only see CMMa politicians around elections

and they report having made several written requests to the CMMa articulating their issues with no

CMMa response.125 Lastly, the Merchant Association reported that trash containers were regularly

RYHUIORZLQJEHFDXVHWKH&00D¶VWUDVKWUXFNVVRPHWLPHVFDPHRQO\RQFHHYHU\WZRZHHNV

125 2QH0HUFKDQW$VVRFLDWLRQPHPEHUSUHVHQWHGLWDVIROORZV³7KH&00DLVOLNHDIDWKHUWKDWVD\VKH is going to buy you shoes. He leaves you waiting and hoping you will receive the shoes but they never DUULYH´

136

Weak Municipal Presence/Strong Party Influence

$VSUHYLRXVO\VWDWHGWKH0HUFKDQW$VVRFLDWLRQVLQ0DWROD¶VPDLQPDUNHWVZHUHZHDNEHFDXVH

they both did not have access to funds and had no influence over municipal policy. Furthermore,

interviews indicated that levels of corruption in the market and municipal mismanagement may

have discouraged cooperation between Merchant Associations and the CMMa because vendors did

not trust the municipality. The absence of regularized elections of Merchant Association presidents

and the lack of a consistently-imposed fee structure did not provide Merchant Associations the

types of incentives present in Maputo and Dondo to take ownership over market maintenance or fee

collection which directly affected municipal services to vendors in these markets. The municipality

did not have the capacity to either partner with Merchant Associations or fill the gap. The CMMa

lacked the personnel to collect and monitor fees that could be used in a revenue-sharing

DUUDQJHPHQWVLPLODUWRWKDWRI0DSXWR¶VZKLFK created incentives for Merchant Associations to

NHHSWUDFNRIH[SHQVHVDQGFRPSHOYHQGRUVWRSD\7KH'0)¶VLQDELOLW\WRWUDFNWKHQXPEHURI

markets and vendors resulted in a low level of intake and a high level of revenue leakage. Up until

2008, the CMMa had no modern, effective system in place to track fees in its 154 markets.126 Site

visits were irregular, at best, so the DMF conducted little to no verification of vendor population or

fee collection. In two markets, market chiefs provided numbers of vendors and daily intake that did

not match the estimates provided by the DMF.127

Maputo, Beira and Dondo all struggle to fully collect fees and have suboptimal market

infrastructure. The key difference with Matola was that Merchant Associations in markets I studied

there had no partnership with the municipality, resulting in a weak municipal presence in the

markets and the lack of political will to address market chief corruption. One municipal staffer

126 CMMa DMF list of markets and 2007 -2008 fees.

127 In one market, the chief indicated that had 500 vendors and collected only 160 MTn per day. Based on CMMa fee records, that market chief underestimated vendors by half. In another market, the chief indicated that he had 290 vendors and collected only 290 MTn, which again, was highly unlikely if fees are between 4-6 MTn. One DMF staff member indicated that these individuals are generally not able to justify fee collection or resolve problems in the markets.

137

shared that fee collectors were known to regularly skim funds from their collections, especially

over the weekend, when they kept fees collected Friday until turning them in to the municipality on

0RQGD\7KLVOHDNDJHZDVWKHUHVXOWRIWKH&00D¶VSROLWLFDOFRQVLGHUDWLRQRIPDUNHWFKLHIV

Many of these fee collectors are ex-combatentes who fought for FRELIMO during the colonial war

of independence, so CMMa political leadership is extremely hesitant to prosecute this practice,

when these individuals receive so little compensation otherwise.

Matola Trash Collection

Interviewees and focus groups indicated that there were no civil society organizations working

in the area of trash collection. Unlike the other three municipalities, interviewees suggested that

trash collection had not comparably improved in Matola and that corruption was to blame.

According to interviews with trash collectors DQGIRFXVJURXSVWKH&00D¶VWUDVKFROOHFWLRQ

service has not improved since 2000, when the CMMa had no trash containers and piles of trash

festered in the streets.128 Service was erratic and linked to the bairro leadership. One participant

stated that residents have had to chase the trash tractor down the street because they did not stop to

collect the trash.129 Another focus group participant stated that that tractor does not have a

structured route in the neighborhood and tends to pass all the houses on the block, only to stop at

the bairro 6HFUHWDU\¶VKRXVHWRFROOHFWKLVWUDVK2QHIRFXVJURXSSDUWLFLSDQWVKDUHGWKDWWKH

³&00DLVDPDILD´$QRWKHULQWHUYLHZHHLQGLFDWHGWKDWWKH&00DQHHGVWRILQLVKZLWKWKHVRFLDO

128 I conducted two focus groups of 16 residents from Matola A and Matola C. Matola A is the formal neighborhood and Matola C is the informal neighborhood. They are adjacent. One trash collector who had been working with the CMMa since 1980 indicated there had been no improvement at all, the only trash collector who indicated this off all those interviewed. All trash collectors interviewed complained about not having sufficient pay or health (in all cases, trash collectors get paid minimum wage which is DERXW07QPRQWKRU86' +RZHYHU0DWROD¶VLQWHUYLHZHGWUDVKFROOHFWRUVZHUHWKH only ones who admitted that conditions had not measurably improved. With the exception of Maputo, where I observed all the sweepers and trash collectors don protective gear, all the other municipal trash collectors had varying degrees of access to protective clothing.

129 $OO0DWROD$SDUWLFLSDQWVLQGLFDWHGWKDWWKH\OLYHLQ³FOHDQ´RU³VHPL-FOHDQ´FRQGLWLRQV/LWWHU Photographs, Matola A focus group, Matola, November 20, 2008. Another Matola C participant indicates that the tractor does pass, but service is only well-done around election time. Another Matola C participant indicated that people are running out of room to bury trash in their yards and need the CMMa to improve service. Participants, Matola C focus group, Matola, November 20, 2008.

138

discrimination when they pick some areas and not others to be serviced just because some chiefs or

³9,3VOLYHWKHUH´/DFNRIFRQILGHQFHZDVVRKLJKWKDWKDOIRIIRFXVJURXSLQRQHQHLJKERUKRRG

indicated they should stop paying fees, the highest of any case study city. Corruption drove the

provision of municipal services, such as trash collection. Even one municipal staffer admitted that

PXFKGRHVQRWKDSSHQLQ0DWRODEHFDXVHRI³ODFNRISROLWLFDOZLOO´130

Lack of Municipal and Community Capacity

The CMMa lacked basic capacity²specifically, trained personnel and systems²to deal with

WKHPXQLFLSDOLW\¶VJURZWKDQGFRUUHVSRQGLQJIORRGRIWUDVK and this was evident in several

examples. Of the four cases, Matola had the lowest ratio of trash collectors to residents, indicating

that there are not enough sanitation workers to keep up with the trash flow (Census 2007).131

Unlike the other three cases, the CMMa has no audit staff to raise awareness about trash collection

or to fine residents for improperly disposing of trash. This lack of personnel and procedure was so

pronounced that some donors have decided not to work with the CMMa. One donor that works in

solid waste management shared that they tried to include Matola in a technical assistance pilot

SURMHFWEXWDEDQGRQHGWKHSDUWQHUVKLSZKHQWKH&00D¶V lack of staff capacity became an

impediment to the project. Unlike the other three cases, Matola suffers from a dearth of community

organizations that work in the area of trash collection and sanitation and no alternative to

interacting with the municipality.

130 There were inconsistencies in the information gathered about the financial aspects of trash collection. First, municipal staff indicated that the CMMa pays a 15 percent monthly share to EDM for fee collection in the bill; however, the Finance Town Councilor indicated that it was 10 percent because 0DWRODKDGD³VSHFLDOGHDO´6HFRQGthe trash regulations indicate that Matola operates on a fee scale associated with energy use, like Maputo. However, focus group members generally quoted a 31 MTn FKDUJHDQGWKH6DQLWDWLRQ7RZQ&RXQFLORU¶VDVVLVWDQWLQGLFDWHGWKDWLWZDVEHWZHHQDQG MTn/month.

131 Beira has the lowest with 1,621 residents per trash worker, followed by Dondo with 2,556 residents per trash collector, then Maputo with 3,140 residents per trash collector. Matola is sorely understaffed, with 42,213 resident per trash colleFWRU$XWKRU¶VFDOFXODWLRQVXVLQJGDWDIURPPXQLFLSDOVWDII interviews and 2007 Census.

139

Conclusion

)DWWRQVWDWHGWKHUHODWLRQVKLSEHWZHHQFLYLOVRFLHW\DQGWKHVWDWHEHVWLQREVHUYLQJWKDW³WKHUHLV a dialectical interaction between state and civil society. The state is transformed by a changing civil society; civil society is transformed by a changing state. Thus, state and civil society form a

IDEULFRIWLJKWO\LQWHUZRYHQWKUHDGVHYHQLIWKH\KDYHWKHLURZQLQGHSHQGHQWSDWWHUQV´ )DWWRQ

1995, 67). I began this chapter hypothesizing that strong civil society organizations affect municipal performance. I argued that civil society would affect municipal performance because participation enables civil society, either as organizations, such as Merchant Associations and

QHLJKERUKRRGWUDVKJURXSVRUSURFHVVHVVXFKDV'RQGR¶VSDUWLFLSDWRU\EXGJHWSURFHVV or boycotts as was the case with the trash fee in Maputo, exert pressure on municipalities to improve municipal services. In order to examine trash collection and market management services, I studied the organizations and processes that existed in these spaces. I gauged the strength of civil society organizations and processes by their access to and control over financial resources or their financial sustainability and their ability to influence municipal policy. I found that: strong civil society is not a prerequisite for strong municipal performance in the two test areas; the relationship between civil society and the municipality mattered in terms of how civil society was able to assist the municipality in improving services in the two test areas; and, that donors that provided aid which built the capacity of civil society organizations to perform services led to sustainable improvements in trash collection in informal neighborhoods. I also found that Maputo was the only case in which

I observed civil society in these two areas act as an independent political force in ASSOSTI and in the neighborhood organizations that become city contractors. I also found that Maputo was the only case in which I observed Merchant Associations receive a portion of municipal revenue to maintain market infrastructure. Lastly, it was the only case in which a boycott of trash fee payments led to a revision of the trash fees.

140

Maputo, Beira and Dondo all were determined to have high levels of performance in the test areas even though Maputo was the only case where civil society was deemed to be strong. In

'RQGRWKLVZDVWKHUHVXOWRI0D\RU&DPEH]R¶VOHDGHUVKLSLQLQVWLWXWLQJYHUWLFDODFFRXQWDELOLW\

FKDQQHOVWKURXJK'RQGR¶VSDUWLFLSDWRU\EXGJHWSURFHVVDQGWKHGHFHQWUDOL]DWLon of fee collection in the major markets to the Merchant Associations. In Beira, success in trash collection was due to

0D\RU6LPDQJR¶VOHDGHUVKLSDQGWKHUHODWLRQVKLSWKDWWKH0HUFKDQW$VVRFLDWLRQOHDGHUVKLSKDGZLWK the CMB. In Maputo, Merchant Associations and neighborhood associations were strong and they positively affected municipal performance market management and trash collection. However, as previously discussed, the informal vendors lobby ASSOTSI, while strong, was motivated by a desire to preserve its access to informal vendor member payments rather than working to improve vendor access to municipal services.

In Maputo and Dondo, donors provided assistance that built capacity in civil society. In

Maputo, donors provided supplies and assistance in formalizing neighborhood associations to collect trash in the informal neighborhoods. They also worked with the CMM to help these organizations gain access to municipal contracts so that they could consistently extend the

PXQLFLSDOLW\¶VSURYLVLRQof trash collection services to previously inaccessible areas. In Dondo,

WKH$XVWULDQ'HYHORSPHQW&RRSHUDWLRQ¶VLQYHVWPHQWLQ'RQGR¶VSDUWLFLSDWRU\EXGJHWSURFHVVJDYH residents a direct link to the municipalityH[WHQGHGWKHPXQLFLSDOLW\¶VDELOLW\WRJHt citizen feedback by placing CMD town councilors in charge of hearings in appointed neighborhoods and ability to provide input on the kinds of projects they wanted to see funded in their neighborhoods on an annual basis. In Beira, donor assistance focused on providing supplies to neighborhood organizations to clean, sweep and collect trash but this assistance did not result in sustainable organizations extending the reach of municipal services or processes that instituted accountability mechanisms between the residents and the municipality.

In all four cases, institutional factors²specifically, consistent rules such as election terms, fee schedules, and revenue-sharing²played a role in incentivizing ownership and accountability. In

141

Maputo, the Merchant Association president had set terms, fee schedules were set and differed between formal and informal markets. Revenue-sharing, in the form of the 10 percent pay-back, was an additional incentive for Merchant Associations to both perform and act professionally to track funds. While the degree to which funds were tracked in formal and informal markets varied

LQSDUWEHFDXVHRI$66276,¶VLQIOXHQFHDQGOHYHORIHGXFDWLRQDPRQJ0HUFKDQW$VVRFLDWLRQ members, fee funds were used in all markets to make improvements based on the recommendations

RI$VVRFLDWLRQOHDGHUVKLS,Q'RQGRWKH&0'¶VGHFLVLRQWRGHFHQWUDOL]HIHHFROOHFWLRQV²an essential municipal function²to some Merchant Associations resulted in reasonable levels of collections. While both Beira and Matola Merchant Associations had unclear election terms and fee schedules, Beira had strong mayoral leadership, a factor, which I further explore in Chapter 6.

In this chapter, I discussed how some of the microincentives behind revenue-sharing with

Merchant Associations created ownership over market infrastructure maintenance and fee incentives to increase fee collection among vendors in markets like Xipamanine. In the next chapter, I examine the revenue trends across four cases to see how the opportunities decentralization provided through taxation authorities may have affected own-source revenue generation and how revenue source affect municipal performance.

142

Table 4.1. Hypothesis and Findings

Hypothesis: Decentralization creates incentives that motivate municipal government officials to act and opportunities for residents to pressure them.

POLITICAL PRESSURE

Civil Society:

Expected Finding: Strong civil society can pressure municipal government to improve municipal services.

143

Table 4.2. Expected versus Actual Findings

CIVIL SOCIETY M UNICIPAL PERFORM ANCE Expected Actual

M APUTO

Markets Strong. Strong informal lobby challenged CMM and did not lead to improved service for informal vendors. Merchant Associations had high technical capacity, provide municipal services and receive a share of vendor

fees. Strong High Trash Strong. Donors developed neighborhood-based trash Collection collection capacity, which extended municipal capacity to trash to informal communities through municipal contracts.

BEIRA

Markets Active. No substantial challenges to municipality. Merchant Associations act as facilitators for municipality and have assisted in major market moves and renovations.

Trash Strong Active. Neighborhood organizations pick up trash in High Collection informal neighborhoods but activities were contingent on donor funding.

DONDO

Markets Active. No substantive challenges to municipality. In a few markets, Merchant Associations, collected fees on behalf of the municipality.

Trash Strong Active. Neighborhood groups raise awareness but do not High Collection provide services. However, a donor provided substantial assistance towards instituting a participatory budgeting, which has assisted in involving residents in decisions about municipal infrastructure fund allocations.

M ATOLA

Markets Weak. Merchant Associations are weak and ineffective.

Trash Weak Weak. Neighborhood organizations are not active. Low Collection

144

V: REVENUE EFFECTS ON SERVICE DELIVERY

The way I see the role of the municipality is that they charge taxes and we pay. We understand WKHPXQLFLSDOLW\GRHVQRWKDYHDORWRIURRPWRPHVVZLWKWKHSHRSOH¶VZDOOHWV

Maputo Vendor, October 1, 2008

More than lament or traverse the world signing accords of financing and cooperation for programs designed to lead with the national deficit, perhaps it is more important and fundamental that the leaders worry about how to use the resources that are available (emphasis added).

Noé Nhantumbo (Nhantumbo 2008, 21)

Introduction

The purpose of this chapter is to explore how revenue affects municipal performance in four

Mozambican municipalities. I found that own-source revenue grew in the presence of majority shares of unearned income, and that unearned income²specifically central government transfers and donor aid² improved municipal performance in cases with strong mayoral leadership. Specifically, my evidence indicates that own-source revenue (vendor and trash fees) grew in the presence of majority shares of unearned income. Municipal performance in three cases improved in the absence of a tax lobby to pressure it. Large shares of own-source revenue or own-source revenue growth are not a pre-requisite for strong performance. Lastly, unearned income improved municipal performance in cities with strong mayoral leadership. These findings build on theory exploring the relationship between unearned and earned income and government performance by providing evidence that may warrant rethinking of both how much unearned income in practice disincentivizes government performance and accountability and the role of mayoral leadership in managing resources (see Table 5.5).

As discussed in Chapter 2, the fiscal contract literature examines how different types of revenue affect government performance by triggering varying degrees of accountability. One view 145

posits that states that earn income have a high quality of governance because taxpayers expect the government to deliver services and will pay when they see their tax dollars at work (Moore 2008, 34;

Ross 2004; Bräutigam 2008). Consistent with this view is that unearned income, or rents like income resulting from aid or from natural resource wealth, reduces the pressure for the state to bargain with tax payers because the state has an alternative source of income and is not forced to perform (Moore 2008;

Ross 2001). For example, high levels of aid²a type of unearned income²have been associated with declines in quality of governance (Bräutigam and Knack 2004). As I discussed in Chapter 2, I viewed sector-specific aid that I examined at the municipal level as unearned income for two reasons. First, while project-based aid to my four municipal cases is sector assistance and not budget support associated with the discussion of unearned income, it is reported as income in the municipal balance sheet. While municipalities have limited flexibility over these project funds as compared to own-source revenue, municipalities work with donors to determine how to spend the funds (equipment v. salaries/personnel) so they have some discretion. Second, I consider central government transfers in the same way. These transfers have to be spent either on infrastructure or municipal personnel. However, municipalities have some discretion in determining how these funds are allocated.

According to this view, I would expect that earned income would improve municipal performance because the consumer relationship created by fee and tax payers would impose revenue pressure on the municipality to be accountable and provide services. Specifically, I would expect that cases with majority shares of earned income relative to overall receipts would be high performing because the municipality is accountable to fee payers. Alternatively, I would expect that cases with majority shares of unearned income relative to overall receipts would be low performing, because the ability for fee and tax payers to pressure the municipality is low.132 More to the point, I would expect that municipalities with low shares of unearned income²namely, aid and central government transfers²will have low performance (see Table 5.5).

132 While I analyzed fee and tax revenue from 2004 to 2007, I focus on fees because these constitute the largest source of earned income in the four cases. As such, I ZLOOEHUHIHUULQJWR³IHHSD\HUV´ZKHQ discussing the fiscal relationship between fee and tax payers and the state.

146

Mozambique is one of the largest aid recipients in Africa with an average USD$58 per inhabitant, double the sub-Saharan Africa average of USD$26 (Nuvunga 2007,39). As aid levels at the municipal level suggested in Chapter 1, I expected poor municipal performance for all four cases.

However, as I discussed in Chapter 3, I observed variation in service delivery among the four cases, which indicated that, even in an environment with high levels of unearned income, municipalities consistently pick up trash and rehabilitate markets. My findings explored three points of the theoretical debate and provided empirical proof as to why municipal service delivery varied in an environment of high levels of unearned income in ways that require rethinking the current debate.

The first point is that governments respond to tax lobbies that command a dominant share of state revenue because these lobbies can effectively bargain with the state and incentivize better performance. The literature underemphasizes what happens in cases when there is no major tax lobby to bargain with the state.133 In all four cases, the majority share of own-source revenue was a constellation of fees, largely fees from vendors and trash collection. Municipal performance in the two test areas improved absent a major tax lobby in three of the four cases. In the case of Maputo, participation increased as the informal vendor lobby was able to influence municipal policy in its favor and preserve its access to fees from its members.

The second point is that growth in earned income (fees and taxes) should result in better services because governments will be responsive to their fee-SD\LQJ³FXVWRPHUV´,IRXQGWKDWODUJH

(and growing) shares of earned income did not in and of itself result in high municipal performance.

133 As discussed in Chapter 4, in Maputo, the informal vendors association, ASSOSTI was the only case of a major fee lobby that challenged the CMM. It used both revenue and political pressure to achieve its objectives. I discuss it in the previous chapter rather than here for the following reasons. First, while ASSOSTI did boycott fee payments as the episodes in Chapter 4 indicate, ASSOSTI was able to pressure the CMM to change municipal policy on fees because of the political pressure it applied² VSHFLILFDOO\WKH$66267,SUHVLGHQW¶VUHODWLRQVKLSZLWKWKH3UHVLGHQW6HFRQGWKHFLYLOVRFLHW\ literature on the role of civil society organizations in democratization focuses on how strong civil society organizations can lobby the state to pressure for access to more resource, more representation, and better services. I found that ASSOSTI, while an effective lobby, was less interested in representing its constituents and more interested in entrenching its position and access to vendors who paid them membership fees.

147

Additionally, I also found a case of a high performing municipality with the lowest share of own-source revenue, indicating that high levels of earned income alone is insufficient in explaining government performance. Own-source revenue increased in relation to overall receipts, regardless of municipal performance, across all four cases. In Matola, the case in which municipal performance in the two test areas was found to be low, own-source revenue grew on average 26 percent from 2004 to 2007, the fastest of all four cases. In Dondo, where own-source revenue to overall receipts was the lowest (22 percent, on average from 2004 to 2007) municipal performance in trash collection and vendor services were high.

The third point is that governments that rely primarily on unearned income²specifically in this case central government transfers and aid²have little incentive to provide services so government performance is expected to be weak. I found that the literature both underplays the positive role of unearned income in developing the capacity and creates a tension related to the study of subnational revenue sources. Unearned income can develop the institutions of revenue-collection and provide a framework for coordinating all sources of revenue²earned and unearned²to improve services. In cases where the municipality coordinated with donors, municipal performance improved. The World

%DQN¶V3520$3872SURMHFWLQ0DSXWRZDVDQH[DPSOHRIVXFKFRRUGLQDWLRQ7KH$XVWULDQ

'HYHORSPHQW&RRSHUDWLRQ¶VLQYHVWPHQWLQ'RQGR¶VSDUWLFLSDWRU\EXGJHWSURFHVVZDVDQRWKHUVXFK example. Furthermore, donor investment in the three high-performing municipalities may actually be a decisive factor in the increases in own-source revenue (trash fees). Dondo, Maputo and Beira received infrastructure investment from the World Bank for solid waste management and fees increased from 5 percent in Dondo to as high as 23 percent in Maputo from 2004 to 2007. Matola did not receive similar infrastructure investment, which may have contributed to why trash fees actually declined 13 percent in the same time period.

Lastly, my research yielded two additional findings. First, as expected, the two cases in which earned income was a majority share of municipal revenue ±Beira and Maputo±saw improvements in municipal service. Earned income constituted at least half of overall receipts in Beira and Maputo.

148

Central government transfers (63 percent) as a share of overall revenue were highest in Matola, while own-source revenue and aid was low as an overall share of revenue (35 percent and 9 percent, respectively). Second, mayoral leadership surfaced as the key difference between the three high performing municipal cases and the low performing municipal case. Leadership in coordinating donor assistance, own-source revenue generation and investment in municipal services was a condition absent in Matola but present in the other three cases. Beira, Maputo and Dondo had strong mayors with technical background, which helped them install teams and mobilize municipal resources. Even donors shied away from Matola because of lack of municipal capacity and leadership, which supports my finding of the importance of leadership in catalyzing municipal performance improvements. I found two examples of this that I discuss in the following section. A critical component of mayoral leadership-- specifically, the ability for a mayor to be effective²is the role of the party politics. Party politics in

Beira provided its mayor the ability to work directly with voters. This dynamic of the effects of politics on municipal performance will be furthered explored in Chapter 6.

In the following sections, I analyze the evidence that supports the findings stated above. First,

I review the overall trends of revenue flows from 2004 to 2007 across the four cases to access the state of revenue Second, I assess how revenue flows in each case²specifically, revenue composition and patterns of growth in earned and unearned income²affect municipal performance by changing the municipal incentives. Then, I conclude the chapter with a review of these findings and a preview of the following chapter on the effects of the political pressure created by elections on mayors to improve municipal performance.

Overall Trends: Revenue Flows

I expected that the high level of unearned income would disincentivize these municipalities from generating earned income; however, the budget analysis indicated that, on average, earned and unearned income grew at approximately the same rates. Receipts across all four cases almost doubled

149

(178 percent) from 2004 to 2007. Growth was driven primarily by own-source revenue (135 percent increase) and transfers (164 percent increase).134 Cross-case comparison showed four trends. First, high levels of unearned income did not appear to dampen the growth of earned income across cases.

Second, no single tax lobby surfaced to exert revenue pressure on the municipality. Third, own-source revenue grew in Matola, where municipal performance was low for the test services and the composition of revenue varied across municipalities. Lastly, Dondo²a case of high municipal performance²had the lowest share of own-source revenue to overall receipts.

Own-source revenue

Earned income grew rapidly across all four cases, as stated above, and, on average constituted almost half of overall receipts, indicating that these four municipalities had incentives to pursue own- source revenue from 2004 to 2007 as reflected in its growth. Averaging 42 percent of overall receipts for the four cases, own-source revenue consisted primarily of fee receipts (receitas não fiscais) and tax receipts (receitas fiscais). Maputo and Beira, two cases rated as high performing municipalities, had the highest average share of earned income, with 60 percent and 49 percent, respectively, of overall revenue constituted by own-source revenue. Matola, rated as a low-performing case, fell in the middle, with an average 35 percent of earned income to overall receipts. Dondo, identified as a high-performing case, had the lowest share, with only 22 percent of overall receipts generated by own-source revenue (see

Table 5.6).

Further review revealed two important insights: that no single fee or tax dominated and fee receipts were the highest source of own-source revenue. On average, the four case study budgets had 30 fee and tax line items (including proceeds from capital sales), with Dondo having the largest number (45 fees and taxes) and Maputo, the smallest (15 fees and taxes). Although no single own-source item

FRQVWLWXWHGDPDMRULW\VKDUHRIRYHUDOOUHFHLSWVIHHVFRQVWLWXWHGWKHOLRQ¶VVKDUHRIRZQ-source revenue

134 Author calculations, CMM Contas de Gerência de 2004, 2005, 2006, and 2007; CMB Contas de Gerência de 2004, 2005, 2006, and 2007; CMD Contas de Gerência de 2004, 2005, 2006, and 2007; CMMa Contas de Gerência de 2004, 2005, 2006, and 2007.

150

at an average of 75 percent, and had shown the most growth, almost doubling on average from 2004 to

2007 so fees were one driver of the overall increases in municipal budgets. 135 Among fees, trash and market fees constituted, on average, 48 percent of total fees, with Maputo having had the highest share

(68 percent) and Dondo the lowest (35 percent) with Beira (41 percent) and Matola (50 percent) in the middle (see Table 5.6 ,GLGQRWH[SHFWWKDW0DWROD¶VVKDUHRIIHHVZRXOGEHVRKLJKJLYHQWKDW0DWROD was a case where trash collection and market services were rated low and pocked with corruption.

Trash and market fees constituted almost half of total fees, implying that this municipality should have felt revenue pressure to provide better services to these fee payers. This is not the case. In contrast,

Dondo had the lowest share of fees as a percentage of income while focus groups suggested services were reliable. I would have expected fees to exert revenue pressure for higher municipal services.

However, my findings suggest another factor may be driving performance.

Transfers and central government dependence

Like own-source revenue, central government transfer growth increased from 2004 to 2007.

While own-source revenue was, on average 41 percent of total receipts across the four cases, central government transfers were still the highest source of income for all four municipalities, constituting on average 49 percent of overall receipts. Matola had the highest share of transfers (63 percent) followed by Dondo (50 percent), Maputo (44 percent) and Beira (42 percent) (see Table 5.6). Consistent with fiscal contract theory, Matola, rated as low-performing, had the highest share of unearned income in the form of transfers (63 percent) of overall receipts). Half the resources of Dondo, Maputo and Beira came from central government transfers so I would expect that, comparatively, these municipalities depend more on earned revenue than unearned revenue.

135 Although fees and taxes are consistently the largest shares of own-source revenue from 2004 to 2007, I included proceeds from sales of municipal property. These amounts for all cases for this time period were nominal. Author calculations, CMM Contas de Gerência de 2004, 2005, 2006, and 2007; CMB Contas de Gerência de 2004, 2005, 2006, and 2007; CMD Contas de Gerência de 2004, 2005, 2006, and 2007; CMMa Contas de Gerência de 2004, 2005, 2006, and 2007.

151

While transfers accounted for such a large percentage of overall receipts, two facts emerged which might explain how the nature of these transfers could encourage growth in own-source revenue and higher municipal performance. First, while one of the two major transfers funds salaries for municipal staff, elected officials in each municipality must be paid from own-source revenue and, as such, these officials might be incentivized to raise revenue with the new decentralization authorities.136

In order to do so, elected officials would be compelled to provide better municipal services in order to justify the increase in fees and taxes.

Second, transfers shifted in character from primarily funding municipal staff, which arguably allows the central government to have more control over local staff, to investing in municipal infrastructure, which can lead to longer-term service improvements. Despite the fact that the municipal subsidy transfers were the largest share of transfers, these transfers have dropped as a share of overall transfers an average of 20 percent from 65 percent in 2004 to 45 percent in 2007 while infrastructure transfers increased 7 percent, on average from 31 percent in 2004 to 38 percent in 2007 as a share of transfers. Not only did central government transfers shift from municipal personnel compensation to infrastructure investments as a share of transfers but municipal staff subsidies transfers shrank as a share of overall receipts. Municipal staff subsidies declined as an average share of overall receipts from 32

136 The two primary central government transfers included in these budgets were the FCA (Fundo de Compensação Autárquica, Municipal Compensation Fund) and the FIIL (Fundo de Investimento de Iniciativas Locais, Local Initiative Investment Fund). Throughout the chapter, when I refer to central government transfers, I mean FIIL and FCA unless otherwise specified. Maputo, Dondo and Matola have either other state transfers or consigned receipts from the state, which are included in the overall transfers calculation. However, I focus on FIIL and FCA transfers for two reasons: first, they are the PDMRULW\VKDUHVRIWUDQVIHUVDQGVHFRQGLQVRPHFDVHVLWDSSHDUVWKDWGRQRUIXQGVDUHLQFOXGHGLQ³RWKHU stDWHWUDQVIHUV´,GLGQRWLQFOXGHWKHVHIXQGVDVWUDQVIHUVEHFDXVH,ZDQWWRH[DPLQHIXQGVRYHUZKLFK the central government has control, ZKLFKZRXOGEHRQO\WKH),,/DQG)&$QRWGRQRUIXQGV%HLUD¶V budgets reflect only the FIIL and FCA as central government transfers.

The FCA funds salaries for municipal staff and FIIL funds infrastructure investments. In 2009, the State Budget allocated USD$27.9 million (MTn837 million) for the FIIL and FCA. Beatrice Reaud and %HUQKDUG:HLPHU³0R]DPELTXH'HFHQWUDOL]DWLRQ$VVHVVPHQW´5HSRUWSUHSDUHGIRUWKH8QLWHG6WDWHV Agency for International Development. Contract: DSD-I-00-04-00227-00, Task Order 15, June 2010: 10.

152

percent in 2004 to 24 percent in 2007 whereas infrastructure investments have grown slightly on average from 15 percent in 2004 to 17 percent across cases for the same time period.

Foreign aid

$FFRUGLQJWR81,&()0R]DPELTXH¶VVKDUHRIIRUHLJQDLGWR*'P (15 percent) is almost double of the Sub-Saharan African average (6-SHUFHQW ³0R]DPELTXHDWD*ODQFH´ $VVXFK, would expect that, inter alia, municipal performance in my cases would be low because unearned income was so high. My analysis suggests two important counter-points to commonly-held notions regarding aid and Mozambique: first, although Mozambique is one of the largest-aid recipients in sub-

Saharan Africa, municipal budgets demonstrated that they received a relatively small share of aid

(Nuvunga 2007,39). Second, as the following sections will show, strong mayoral leadership and aid tied to specific, capacity-building projects were two factors present in the case-study cities with high performance.

Although central government transfers constituted the largest share of municipal revenue, donors have exerted a major, if somewhat uncoordinated, role in municipal development.137 My analysis of over 1,200 projects over an eight-year period concluded that aid to the 33 Mozambican municipalities was estimated at USD$649 million, or an average of USD$77 million annually, a total of

PRUHWKDQHLJKWWLPHV0DSXWR¶VEXGJHWWKHODUJHVWPXQLFLSDOEXGJHWRIDOOPXQLFLSDOLWLHV138 While aggregate level data indicated that donors committed high levels of aid to municipalities, case study

137 Over the past decade, 14 donors have provided varying levels of support for either decentralization or deconcentration initiatives in Mozambique, primarily through stand-alone projects. A recent effort to coordinate aid for decentralization resulted in the Joint Program to 13 municipalities in Central and Northern Mozambique (P 13). Funded at USD$17 million from 2008-2010, the P 13 is funded by Austria, Denmark and Switzerland is the first large-scale effort to make assistance coherent over several municipalities. See Reaud and Weimer, 27.

138 Author calculations, ODAmoz database, 2004 to 2011. It should be noted that at the time this analysis was conducted, donors were not disaggregating aid by municipality so I estimated the amount of aid as an even share of aid per project. Reporting has since been modified to enable donors to report percentage shares that will be awarded to municipalities and districts. However, since some projects have since closed out and data is no longer available, I continue to use my original analysis.

153

municipal budgets actually report receiving far less. Second, municipalities may not have control over the aid they receive. Although aid dwarfs municipal receipts when measured through donor reports, municipal budgets (Contas de Gerência) point in a different direction.139 From 2004 to 2007, donor disbursement reports indicated that aid to municipalities is on average three times more than overall receipts for these municipalities from 2004 to 2007. However, for the same time period, municipal budgets report far less.

Aid to Beira, Matola and Maputo was low (on average, 9 percent of the total budget for each case) with aid to Dondo averaging 24 percent of total receipts over the same period (see Table 5.6).140

While donors commit funds for projects at the municipal level, donors indicated that funds for these, and other projects, were mainly transferred to a central government line ministry and then funds were released by that ministry to the municipality. So, although large amounts of aid are disbursed by the donor to the central government, those amounts then need to be transferred from the line agency to the municipality, which PLJKWDFFRXQWIRUWKHYDULDWLRQLQDLGDPRXQWV'HVSLWH0R]DPELTXH¶V international reputation as one of the highest aid recipients in Africa, case study budgets indicate that

139 When I reference donor reports, I mean reports of aid disbursements as reported in the most comprehensive, public database available, ODAmoz. It is accessible at www.odamoz.org.mz.

140 Aid transfer arrangements differ from case to case based on the terms of the agreements, transfers and types of aid provided. This makes across-case comparison a challenge. First, aid organizations provide in-kind donations, such as desks, computers, or air conditioners, in cases like Matola (chart 21) and Dondo (chart 17). They are itemized and a value is assigned for this donation. In the case of Dondo, those values are then aggregated into the capital benefits (bens de capital) budget line (2.1) under expenditures (chart 5). However, the donations line (2.4) under receipts also lists donations. It is unclear why these values are accounted differently, under receipts and expenditures, instead of one line item under receipts. Furthermore, capital benefits are not always itemized so it is not clear if other in- kind donations from donors are included in this line item. Second, there are two budget line items for donations (1.5 and 2.4) and it remains unclear why. Third, municipalities report donations differently. For example, Beira lists its aid (UNICEF and World Bank projects) as does Dondo (PADM, PDM and UNICEF) under receipts as donations. However, Maputo reports PDM as a transfer from other public entities (transferências de capital de outras entidades públicas-PDM) (2005 Conta, chart 5) under receipts, with donations line 1.5 and 2.4 remaining blank. In 2007, PROMAPUTO is similar listed (lines 2.3.1.2 and 2.3.1.3). Furthermore, PDM funds are also listed under capital expenditures under capital benefits (line 2.1.3) making tracking aid within the municipal budget a challenge.

154

municipalities receive a relatively small share of aid, averaging 13 percent from 2004 to 2007, suggesting the aid investments are limited and project-specific (see Table 5.6).141

Although case study budgets revealed less aid than what was reported in donor disbursement reports, I submit that municipalities, which had control over funds would be better able to be responsive and provide services. So, even in an environment of high unearned income, the aid projects I examined provided sector-specific assistance in which municipalities exerted some control over this income and, while not being directly accountable to fee payers, they were accountable to donors. In order to capture this dynamic, I constructed a category entitled aid autonomy. I measured aid autonomy by examining the amount of aid that goes to municipalities directly versus the amount that is awarded to the central government on behalf of a municipality (see Table 5.7). 142 All aid from 2004 to 2007 for Dondo and

Matola goes through the central government.143 However, Maputo does have a relatively high degree of aid autonomy, with 20 percent of aid going directly to the municipality followed by Beira, which receives 6 percent of its aid directly (see Table 5.7). 144

Of these aid projects, I found two projects that provided assistance to the three high-performing municipalities: TKH:RUOG%DQN¶V0XQLFLSDO'HYHORSPHQW3URJUDP 3'0 DQGWKH$XVWULDQ

'HYHORSPHQW&RRSHUDWLRQ¶VIXQGLQJRI'RQGR¶VSDUWLFLSDWRU\EXGJHWSURFHVV3'0ZDVDYDLODEOHWR all four municipalities; however, Matola did not receive any project funds, meaning that it did not receive funding to purchase trash trucks or make major renovations to markets, which may explain why

141 CMM Contas de Gerência de 2004, 2005, 2006, and 2007; CMB Contas de Gerência de 2004, 2005, 2006, and 2007; CMD Contas de Gerência de 2004, 2005, 2006, and 2007; CMMa Contas de Gerência de 2004, 2005, 2006, and 2007.

142 2'$PR]SURMHFWGRFXPHQWVOLVWWKH³*RYHUQPHQW&RXQWHUSDUW´RU³&KDQQHORI'HOLYHU\´RQZhich I rely to determine how the funds are being disbursed from the donor to the recipients. Channel of GHOLYHU\LVGHILQHGDV³WKHLPSOHPHQWLQJDJHQF\LQFKDUJHRILPSOHPHQWLQJWKHSURMHFW´ODADATA glossary available at www.ODAmoz.org.mz, accessed on July 25, 2010.

143 Aid examined from 2004 to 2007 for these case studies is based on review of projects in the ODAmoz database. Percentages are based on project totals as reported in ODAmoz from 2007 to 2010 to receipts municipalities reported in the Contas de Gerência from 2004 to 2007 for each case.

144 Ibid.

155

trash collection is observed to be comparatively low in this municipality. PDM was a six-year, USD$33 million World Bank-funded project designed to assist the Mozambican government and the 33 newly- designated municipalities to implement decentralization through legal and institutional reform, municipal capacity building and municipal grants. As a wide-ranging capacity-building project, municipal aid autonomy was low because the central government was designated as the channel of delivery. Although all four municipalities participated in the PDM, Matola was the only municipality that was not included in the municipal grants funds that helped Beira, Maputo and Dondo purchased

WUDVKFROOHFWLRQHTXLSPHQWDQGYHKLFOHVDQGKHOSLPSURYLQJPDUNHWLQIUDVWUXFWXUH ³,PSOHPHQWDWLRQ

&RPSOHWLRQ´, 2-3,12,35).145 Dondo received a separate technical assistance award (PADM) from the Austrian Development Cooperation. These funds, while not directly awarded to Dondo, funded its participatory budget process, which assisted the municipality in allocating infrastructure funding.

In conclusion, across cases, two facts surfaced about revenue composition and type, which challenge assumptions about earned and unearned income and government performance. First, earned and unearned income grew at almost the same rates from 2004 to 2007, which suggests that the presence of high levels of unearned income did not hamper the growth of earned income in the test cases. Two factors might explain this growth. Donors invested in capacity building and infrastructure projects in markets and trash collection that may have enabled municipalities to increase fee collection.

Second, while transfers fund municipal staff, elected officials must be paid from own-source revenue, which might incentivize own-source revenue generation and better municipal performance. My research found that elected officials where motivated to increase own-source revenue because it increased their autonomy over resources. Even though Mozambique is one of the largest aid recipients in sub-Saharan

Africa, municipal budgets showed relatively low aid receipt compared to overall receipts. These factors indicate the ZD\WKDWPD\RU¶Vused own-source revenue, rather than the quantity of that revenue, of

145 Author calculations based on actual disbursements of each of the three components. Please see, :RUOG%DQN³,PSOHPHQWDWLRQ&RPSOHWLRQ and Results Report (IDA-35490 IDA 3549A) on a Credit in the Amount of US$33.60 million (SDR 26.70 million credit) to Republic of Mozambique for the 0XQLFLSDO'HYHORSPHQW3URMHFW´5HSRUW1R,&5$)7:DWHUDQG8UEDQ6RXWKHUQ$IULFD Country Cluster 2, AFRICA, August 27, 2007.

156

unearned income has a greater effect on municipal performance and earned income than previously thought.

Second, a variety of fees and taxes (on average 30 fees and taxes per municipality) fed own- source revenue, meaning that no single lobby dominated. While no single lobby dominated, fees were the highest source of own-source revenue and trash and market fees constituted on average 75 percent of overall fees. Unexpectedly, own-source revenue grew fastest in the municipality that rated low on performance. This evidence suggests that factors other than the growth or share of earned income may influence municipal performance. In the following sections, I explore in greater depth how earned income and unearned income within each case affect municipal performance to understand the positive role of unearned income.

Across time, within case

Maputo

M APUTO 2004 2005 2006 2007146 (USD) Revenue Autonomy (Own Source 59% 56% 59% 66% Revenue to Overall Receipts) Largest share of fee revenue Trash fee Trash fee Trash fee Trash fee $726,090 $812,851 $882,829 $1,464,728 Transfers to overall receipts 41% 44% 41% 34% Table 5.1: Maputo, CMM Contas de Gerência de 2004, 2005, 2006, 2007, charts 5 and 6. Maputo is the only case, which has had additional state transfers from 2004 to 2007, which is why the FCA and FIIL do not add to transfers as a share of overall receipts. Exchange rate used USD$1:26.5 MTn.

146 The PROMAPUTO transfer is reflected in transfers, causing them to jump relative to overall receipts from 41 percent in 2004 to 49 percent in 2007. Because PROMAPUTO is such a large FRPSRQHQWRI0DSXWR¶VEXGJHWLQFRPSDULVon to aid projects in the other cases, I removed the PROMAPUTO transfer of USD $2.8 million in 2007 to be better able to isolate the own-source revenue to see how it compared to central government transfers. With the PROMAPUTO transfer, revenue autonomy is lower (51 percent of overall receipts is own-source), transfers are higher (49 percent of overall receipts are transfers), and FIIL and FCA transfers constitute a smaller share of overall receipts (9 percent and 17 percent, respectively).

157

I expected earned income to drive municipal performance in Maputo. I found that, central government transfers as a share of overall revenue shrank from 41 percent in 2004 to 34 percent in 2007

(see Table 5.1). Of all four cases, Maputo had the highest level of earned income as a ratio of overall receipts from 2004 to 2007, growing from 59 percent to 66 percent. The largest shares of this earned income (29 percent) were fees²specifically trash and vendor fees, with the highest growth in trash fees.

:KLOH0DSXWR¶VRZQ-source revenue growth was primarily in taxes, not fees, fees constituted the largest share of own-source revenue, with trash fees constituting the majority share of fee revenue (41 percent on average) and vHQGRUIHHVFRPPDQGLQJWKHVHFRQGODUJHVWVKDUH SHUFHQW ³0R]DPELTXHUHFHLYHV

86'PLOOLRQ´ 147 Growth in fees was primarily driven by growth in trash and vendor fees, which comprise the majority (68 percent) of fee revenue. Trash fees, which form the largest share of fee revenues, doubled from 2004 to 2007 from USD$726 thousand to USD$1.4 million, reflecting the largest and fastest growing fee among all fees. While fees fed service improvements, trash fees have increased due to capacity building and infrastructure investments made by the GTZ, through its

AGRESU program and PROMAPUTO. I specifically found that strategic donor assistance in the form

RI$*5(68¶VWHFKQLFDODVVLVWDQFHRQUHYLVLRQRIWKHWUDVKIHHVFDOH0DSXWRDGRSWHGIRUWUDVKcollection helped set fee levels individuals paid and these increases are seen in the growth of trash fees.

Beira BEIRA 2004 2005 2006 2007 (USD) Revenue Autonomy (own- 48% 52% 42% 55% source to overall receipts) Largest share of fee revenue Vendor fees Vendor fees Vendor fees Trash fees $273,129 $337,100 $364,171 $378,367 Transfers to overall receipts 41% 46% 39% 42%

Table 5.2: Source: CMB Contas de Gerência de 2004, 2005, 2006, and 2007, charts 5 and 8. Beira is the only case has no state transfers or consigned state receipts. Exchange rate used USD$1:26.5 MTn. 147 CMM Contas de Gerência de 2004, 2005, 2006, and 2007. PROMAPUTO was extended under the

International Development Association terms, with a small service charge and 40 year period of maturity, which can be extended by another ten years. PROMAPUTO was approved January 25, 2007 DVSHUWKHSURMHFWGRFXPHQWHQWLWOHG³3520$38720DSXWR0XQLFLSDO'HYHORSPHQW3URJUDP´ Wor l d Bank Project ID P096332, dated January 5, 2007. Found on www.worldbank.org

158

Like Maputo, I expected earned income to drive municipal performance in Beira. I found that Beira had the second highest level of own-source revenue to receipts, which grew from 48 percent in 2004 to 55

SHUFHQWLQ2IWRWDOIHHVSHUFHQWZHUHIURPWUDVKDQGYHQGRUIHHV%HLUD¶VRZQ-source revenue as a share of overall receipts exceeded transfers (see Table 5.2). While own-source revenue fed services, own-source revenue grew because of the improvement in services instigated by both Mayor

Simango and some donor projects.

/LNHDOOFDVHV%HLUD¶VEXGJHWincreased from 2004 to 2007, growing by 63 percent to USD$5.2 million by 2007. While both transfers and own-source revenue explained this growth, own-source revenue to overall receipts outpaced growth in transfers, increasing 7 percent from 48 percent in 2004 to

SHUFHQWLQ*URZWKLQ%HLUD¶VRZQ-source revenue was driven by three fees: land use permits, vendor fees and trash fees.148 Trash fees doubled from 2004 to 2007, with that growth primarily taking place between 2006 and 2007 when fees grew from USD$179 thousand in 2006 to USD $378 thousand

LQ6HFRQGWKH&0%¶VODUJHVWSURSRUWLRQ URXJKO\RQH-quarter) of fee collection was vendor fees from 2004 to 2006, with trash fees surpassing vendor fees by 2007 as the largest share of overall fee receipts.

Similar to Maputo and Matola, donor investments in Beira constituted a four-year average of 9 percent of the CMB¶VWRWDOEXGJHW VHH7DEOH5.6). 149 These aid investments were driven by both the

:RUOG%DQN¶V3'0SURMHFW 86'WKRXVDQG LQZKLFKDOOIRXUPXQLFLSDOLWLHVSDUWLFLSDWHGDQG

UNICEF, a water-sewer improvement project (USD$490 thousand). 150 Like Maputo and Dondo, the

148 Trash fees grew 111 percent increase from USD $179 thousand to USD$378 thousand and land use permit fees grew 153 percent increase from USD$72 thousand toUSD$183 thousand.

149 Author calculations based on CMB Conta de Gerência de 2006, Chart 6.

150 Central JRYHUQPHQWWUDQVIHUVZHUHPRUHUHOLDEOHWKDQGRQRULQYHVWPHQW7KH&0%¶VEXGJHW projections from 2004 to 2007 anticipate a much higher level of donor investment than what is reported as transferred by the municipality. The 2005 Beira budget projected a donor commitment, which was 8 times higher than what the CMB actually received. This pattern continued in 2006. Budgeted donor commitments were almost 3 and half times higher than what the CMB actually received. However, 159

World Bank project provided funds for trash equipment and market improvements. However, although donor projects play a role in improving CMB service, these projects did not result in a comprehensive, overarching investment plan as was the caVHZLWK0DSXWR¶V3520$3872SURMHFWDQGWKHUHIRUHZHUH more limited in scope and long-term impact. Furthermore, interviews and focus groups indicated that

0D\RU6LPDQJR¶VDELOLW\WRUHDOORFDWHPXQLFLSDOVRXUFHVTXLFNO\UDWKHUWKDQGRQRULQYHVWPHQWVVSXrred an immediate improvement in trash collection.

Dondo

DONDO 2004 2005 2006 2007 (USD) Revenue Autonomy 28% 22% 19% 18% (own-source to overall receipts) Largest share of fee * Maintenance of Maintenance of Use of municipal revenue gardens and gardens and property markets markets $62,834 $31,598 $37,187 Transfers to overall 40% 48% 42% 69% receipts Table 5.3: Source: CMD Contas de Gerência de 2004, 2005, 2006, and 2007, charts 5 and 10. Dondo has additional state transfers from 2006 and 2007, which are included in the transfer calculations. Data for 2004 was not disaggregated. Exchange rate USD$1:26.5MTn.

I expected high levels of earned income to drive municipal performance in Dondo (CMD) because I observed high performance in one of my two test areas, trash collection, as discussed in

Chapter 3, and I posited that municipal cases with high shares of earned in come will have high performance. However, I found that Dondo had the lowest level of revenue autonomy (22 percent own- source to total receipts) and the highest level of aid dependence (25 percent) (see Table 5.3).

Furthermore, own-source revenue as a share of overall revenue actually shrank in Dondo from 28

examining FIIL and FCA transfers for the same time period revealed that the amount of transfers the CMB included in their planned receipts was actually transferred. Author calculations based on CMB, Chart 5, Conta de Gerência de 2005 and Conta de Gerência de 2006; &0%³0DSD&RPSDUDWLYD(Qtre D5HFHLWD2UoDPHQWDGDH&REUDGD'XUDQWHD*HUrQFLD´Conta de Gerência de 2005, Chart 10. This may have been the case with other municipalities; however, I was unable to find this particular level of detail relating to budgeted donor commitments in the other three case studies.

160

percent in 2004 to 18 percent in 2007. This revenue profile²high dependence on unearned income, specifically, transfers and aid ²when matched with high performance in the two test areas suggests that unearned income²specifically, aid²KDVKDGDSRVLWLYHUROHLQLPSURYLQJ'RQGR¶VPXQLFLSDO performance.

'RQGR¶VEXGJHWIURPWRJUHZ by 78 percent to USD$755 thousand, the largest growth of all four cities. The dynamics of revenue growth detected in the other three cases²increases in own-source revenue and central government transfers²were also at work in Dondo. Like all cases,

'RQGR¶V own-VRXUFHUHYHQXHZDVFRPSULVHGSULPDULO\RIIHHVDOWKRXJKLQ'RQGR¶VFDVHRZQ-source revenue on average was almost exclusively fees (90 percent from 2004 to 2007). On average, trash fees

UHSUHVHQWHGWKHVPDOOHVWVKDUHRI'RQGR¶VRZQ-source receipts while market fee were the largest share of all the cases (see Table 5.6).

Aid and central government transfers constituted the majority (75 percent on average) of the

&0'¶VEXGJHWIURPWR2IWKHIRXUFDVHVWKH&0'UHOLHGPRVWKHDYLO\RQFHQWUDl government transfers (see Table 5.6). From 2004 to 2007, central government transfers grew twice as fast as own-source revenue. From 2004 to 2007, central government transfers on average tripled compared to own-source revenue, which grew 1.5 times. Not only are central government transfers growing twice as fast as own-source revenue, transfers outpaced own-source revenue growth, which

ZDVGHFOLQLQJDVDVKDUHRIRYHUDOOUHFHLSWV7KH&0'¶VEXGJHWVVXJJHVWHGWKDWLQIUDVWUXFWXUH improvements through WorOG%DQN¶V3'0DQG$XVWULDQ&RRSHUDWLRQ¶V3$'0ZHUHEHKLQGWKLVJURZWK in aid. Although I expected that such low levels of own-source revenue would lead to poor service, as I will discuss in the following section, project-specific aid and transfers led to better municipal service

EHFDXVHWKH\IXQGHG'RQGR¶VSDUWLFLSDWRU\EXGJHWSURFHVVVHWXSZLWKWKHVXSSRUWRIWKH$XVWULDQ

Development Corporation, and subsequent investments in market infrastructure.

161

Matola

M ATOLA 2004 2005 2006 2007 (USD) Revenue Autonomy (own- 32% 29% 40% 40% source to overall receipts) Largest share of fee revenue Trash fee Trash fee Trash fee Trash fee $71,051 $182,878 $239,139 $270,969 Transfers to overall receipts 73% 72% 60% 60% Table 5.4: Source: CMMa Contas de Gerência de 2004, 2005, 2006, and 2007, charts 5 and 8. Matola has consigned receipts from the state for 2004 and 2005, which are included in the transfer calculations. Exchange rate USD$1:26.5MTn.

As I previously stated, I hypothesize that high levels of unearned income would result in poor performance. What I found in Matola both confirmed and provided nuance to the fiscal contract debate.

Unearned income constituted the majority revenue in Matola (60 percent in 2007, see Table 5.4). From

2004 to 2007, own-source revenue grew 8 percent from 32 percent in 2004 to 40 percent in 2007, a finding I did not expect given the majority shares of unearned income. Fees constituted the majority of own-source revenue (77 percent) of which trash fees represented the largest share. Although trash fees were the largest share of fee revenue, a spike in own-source revenue from 2005 to 2006 was attributable

ODUJHO\WRDGRXEOLQJRI³QRQ-VSHFLILHGIHHV´

Focus groups indicated that services were selectively provided to party officials and vendor fees were collected under coercive terms. While Matola received the same share of aid as the Maputo and Beira, the conditions that were present in Maputo, Beira and even Dondo²coordinated aid investment²ZHUHQRWSUHVHQWLQ0DWROD)XUWKHUPRUHWKHLQWHUUXSWHGWHQXUHRI0DWROD¶VSRSXODUPD\RU may explain the vacuum in leadership to coordinate and drive use of all revenue sources. This suggests two conditions were present in municipalities with high performance: coordinated donor investment and/or mayoral leadership. Coordinate aid±while unearned income²can provide a valuable counterbalance to central government transfers. The evidence has also indicated that mayoral leadership has played a key role in mobilizing what kind of resources.

162

Consistent with aid flows in the 2004 to 2007 budgets of Maputo and Beira, aid plays a comparatively small role in Matola, averaging 9 percent from 2004 to 2007. With real growth in own- source revenue as a share of overall receipts, like Beira, I would have expected that services would have improved as a result of increased investment. Trash fees are the highest of the four municipalities averaging 38 percent of fees which might indicate that the CMMa is providing a high level of service because: a) the CMMa has the level of professionalization and capacity to collect fees and b) residents are satisfied enough with service that they are willing to pay the CMMa for them. The performance analysis that follows tells a different story.

Performance Analysis

The preceding section reviewed both earned and unearned income as a share of overall resources for the four cases from 2004 to 2007 and yielded two insights into the accountability relationship between government and its tax payers. First, earned income across all cases even in the presence of majority shares of unearned income²largely central government transfers²grew. This own-source revenue growth was largely in a constellation of fees, which meant that no single tax lobby dominated. I found that donors invested in activities that permitted all municipalities to collect more fees. While I would have expected that earned income would drive performance what I found was that unearned income in the form of strategic donor investment provided improvements in service and capacity to collect fees. Second, I found that revenue diversification and how revenue was coordinated²whether through coordinated plans, donor-funded participatory processes or mayoral leadership²were conditions present in three high performing municipalities and absent in the low- performing municipality. This suggests that accountability to donors and the ability for the municipality to exert some control over desiJQDQGLPSOHPHQWDWLRQRISURMHFWDVVLVWDQFHDVH[HPSOLILHGLQ0DSXWR¶V

3520$3872SURJUDPDQG'RQGR¶VSDUWLFLSDWRU\EXGJHWSURFHVVIRVWHUHGDFFRXQWDELOLW\ to fee payers because mayors in both cases took a lead role in the design and implementation of that assistance in an

163

effort to be re-elected.. More specifically, while municipalities were compelled to be accountable to donors in the investment of aid and meeting conditions for implementing aid-related projects, donors

FUHDWHG³SUR[\DFFRXQWDELOLW\´WRIee payers that had previously not been in place.

Maputo

In Maputo (CMM), unearned income²specifically donor assistance in the form of the World

%DQN¶V3'0DQG*7=¶V$*5(68SURMHFW²provided the CMM the targeted technical assistance it needed to both build municipal capacity to collect fees and calibrate fee payers expectations regarding payment and service, fortifying the fiscal accountability between the two. Evidence from interviews and focus groups indicated that this targeted, municipally-directed donor technical assistance and

PXQLFLSDOOHDGHUVKLSFRPELQHGWRLPSURYHERWKWKH&00¶VSHUIRUPDQFHDQGIHHFROOHFWLRQ

0D\RU&RPLFKH¶VDSSURDFKLQ0DSXWRZDVDQH[DPSOHRIKRZPD\RUDOOHDGHUVKLSVWHHUHG resources to improve municipal performance. Elected LQ0D\RU&RPLFKHXVKHUHGLQWKH³RSHQ

JRYHUQPHQW´ governação aberta) in which he regularly held forums every 2-3 months to get citizen

IHHGEDFNLQHDFKRI0DSXWR¶VQHLJKERUKRRGV+HSURIHVVLRQDOL]HGPXQLFLSDOJRYHUQPHQWDQGZHHGHG out ghost employees. He hired competent qualified people, like the town councilors for sanitation, a

GRFWRUDQGIRUPHUGHDQRI0R]DPELTXH¶VODUJHVWPHGLFDOVFKRRODQGPDUNHWVDEXVLQHVVPDQDQG merchant. The town councilor for sanitation shared how there were not even desks and chairs when he

ZDONHGLQWRWKHRIILFHDIWHUKLVDSSRLQWPHQWLQ³:HGLGQ¶WKDYHPRQH\HTXLSPHQWRUVWDII´WKH town councilor emphasized.151 One NGO specified that an example of how his leadership came to fruition was trash collectiRQ³7KHUHZDVDGLIIHUHQFHEHWZHHQSDVWPDQDJHPHQWDQGFXUUHQW

PDQDJHPHQW´KHVWDWHV³WUDVKFROOHFWLRQZDVUHDOO\EDG´152 ³+H &RPLFKH ZHHGHGRXWWKH

151 Author interview, Town Councilor of Sanitation and Salubrity, Maputo City Hall, October 7, 2008.

152 Author interview, Dr. João Pereira, Mecanismo para Apoio da Sociedade Civil, September 18, 2008, Maputo.

164

FRUUXSWLRQFKDQJHGWKHLQVWLWXWLRQDOFXOWXUHDQGV\VWHPDWL]HGDFFRXQWDELOLW\´0$6&VWDIIVKDUHG153

By putting together a qualified executive team with whom he met daily, Dr. Comiche instituted a culture of accountability within the CMM. Town councilors were measured by their ability to set up functions within their departments to deliver services and collect fees and they reported daily on their progress (see Chapter 6 for more detail).

One example is the work that the Town Councilor for Sanitation did to improve trash collections services and fee intake. The CMM pursued coordinated donor support to improve service

DQGIHHFROOHFWLRQ$OWKRXJK3520$3872 VHH&KDSWHU ZDVWKHFXOPLQDWLRQRI\HDUV¶RI stakeholder meetings and provided an overarching framework and financing for equipment to improve trash collection among other services, it was the technical assistance provided by the Geselleschaft für

Technische Zusammenarbeit (GTZ) to the Sanitation Department, which directly impacted trash fee

FROOHFWLRQVIURPWR7KH*7=¶V$*5(68SURMHFWEHJDQLQRIIHULQJWHFKQLFDO assistance to the CMM. The CMM is named as the government counterpart of this project meaning that the GTZ through AGRESU has a direct relationship with the CMM instead of a line government agency, as was the case with donor assistance to Dondo, Beira and Matola.

This close relationship permitted AGRESU to act more nimbly as technical advisor to the

CMM which helped them to better collect fees by conducting studies of market sensitivities to fee increases and working closely with the CMM and the fee payers to introduce a scaled fee schedule, a condition unique to the Maputo case. The CMM introduced a flat fee of 20 MTn/month for all city residents without consultation or public outreach in 2002 (Kruks-Wisner 2006,39). Coupled with no visible improvement to trash collection, residents simply refused to pay the fee.154 Protests took place in both formal and informal neighborhoods, with residents blocking major intersections and the road to the airport. In response, the CMM, with the assistance of AGRESU, undertook two responses. First,

153 Ibid.

154 Joaquim Stretz, AGRESU interview in Maputo at the Feira Ambiental, September 21, 2008.

165

the CMM conducted a public information campaign to explain the use of the fees. Second, in 2003, the

CMM began experimenting with a scaled fee for domestic and commercial producers to replace the flat fee due to boycotts of fee payment. However, it was not until 2004 under the Comiche Administration,

WKDW$*5(68WHFKQLFDODVVLVWDQFHZDVIXOO\XWLOL]HG:LWK$*5(68¶VDVVLVWDQFHWKH&00 conducted numerous studies that enabled the CMM to arrive at a scaled fee schedule based on electricity usage. Maputo is the only municipality with scaled fees in place, which focus groups understood and felt was fairer than the previous flat fee, even if they felt the CMM should be providing more service. The CMM had been able to use this assistance to formulate what the Sanitation town

FRXQFLORUFDOOVKLV³ELEOH´²a general strategic plan which includes how the CMM will eventually phase out PROMAPUTO funds and increase fees to make solid waste management sustainable, based on a phased increase to trash fees informed by AGRESU studies. As a result, trash fees have consistently grown, doubling since 2004, constituting an average of 41 percent of own-source revenue. Among cases, this approach was unique. The combination of mayoral leadership to inculcate a culture of accountability in municipal management combined with targeted donor technical assistance led to an increase in trash fees that fed the overall growth of own-source revenue for Maputo from 2004 to 2007.

Beira

What set Beira apart from the other cases is that Beira receives the smallest average share of central government transfers relative to overall receipts which suggests that Beira was less dependent than other cases on central government transfers as a result of own-source revenue growth. Trash fees grew between 2006 and 2007, surpassing vendor fees. The revenue picture, in combination with results of focus groups, suggests that the fee increase followed a visible improvement in trash collection, which began in 2005 with the election of Mayor Simango. These three factors-low comparative central government transfers, majority sources of own-source revenue that have grown from 2004 to 2007, and mayoral leadership²combined to improve service in Beira. In the absence of pressure from a tax lobby, as was the case across all cases, and targeted donor capacity building, as was the case in Dondo

DQG0DSXWR0D\RU6LPDQJR¶VOHDGHUVKLSSURYLGHGWKHDFFRXQWDELOLW\WRPRWLYDWHLPSURYHPHQWVLQ

166

trash collection and vendor services. As will be discussed iQ&KDSWHU%HLUD¶VVWDWXVDVDQ

³RSSRVLWLRQ´PXQLFLSDOLW\FRPELQHGZLWKDSRSXODUPD\RUSURYLGHGWKHSROLWLFDOSUHVVXUHWRJHQHUDWH improvements in municipal services.

Donors have played a substantial role in assisting Beira in improving services in the areas of market infrastructure and solid waste management since decentralization. Like the other cases, Beira

SDUWLFLSDWHGLQWKH:RUOG%DQN¶V3'0SURMHFW'XULQJWKHOLIHRIWKHSURMHFWWKH&0%UHFHLYHG

USD$2.5 million, which included USD$478 thousand for solid waste equipment and USD$744

WKRXVDQGIRUPDUNHWVDQGRWKHULQYHVWPHQWV ³,PSOHPHQWDWLRQ&RPSOHWLRQ´ )URPWR the CMB received approximately USD$830 thousand for PDM activities, which means that 68 percent of project funds were disbursed between 2002 and 2004, during the first mayoral mandate prior to the election of Mayor Simango (Ibid, 22).155

:KLOHWKHPDMRULW\RIIXQGVZHUHGLVEXUVHGSULRUWR6LPDQJR¶VPDQGDWHIRFXVJURXSV indicated that trash collection did not actually start to measurably improve until 2004, suggesting that

Simango was able to mobilize resources and staff to do what previous mayors were unable to do: pick up the trash and move informal vendors to a formal market. Focus groups indicate that trash collection

GLGQRW³WDNHRII´XQWLODIWHU2QHSDUWLFLSDQWVWDWHG³WKHFLW\LVZRUNLQJODWHO\,WLVHYLGHQW´

³7RGD\\RXGRQ¶WVHHWUDVK´DQRWKHUSDUWLFLSDQWVKDUHG156 Furthermore, major rehabilitation of

Maquinino market also occurred after 2004. Focus groups and Merchant Associations indicate that

WKHVHFRXOGEHFUHGLWHGWRWKH³VHULRXVOHDGHUVKLS´RI0D\RU6LPDQJR,QWKHLUYLHZSUHYLRXV administrations mismanaged funds as they never saw an improvement in service. So, although the aid was available previously, the lack of leadership hindered implementation, a general observation made of

SDUWLFLSDWLQJPXQLFLSDOLWLHVDWWKHFORVHRIWKH:RUOG%DQN¶V3'00D\RU6LPDQJR¶VWHUPEHJDQ

155 The CMB received USD$316 thousand in 2004, USD$80 thousand in 2005, and USD $432 thousand in 2006. Amounts are rounded. See donations line in CMB, Contas de Gerência do Exercicio de 2004, 2005, and 2006

156 Participants, Palmeiras I focus group, March 17, 2009, Beira.

167

February 6, 2004 and within two days, municipal staff were paid back-pay and, within a few months, the piles of trash, which had accumulated in the streets, were collected.157 In 2006, the CMB purchased two container trucks valued at USD$200,000 using own-source revenue, which was reallocated from a

SODQQHGPDUNHWFRQVWUXFWLRQSURMHFWLQ0DUD]p ³1RWLFLDV´ $OWKRXJKWKH,WDOLDQ&RRSHUDWLRQKDV been active in Sofala province, its municipal development program, PADDEL, was not funded until

2007, after these changes occurred. As one focus group participant stated: what was needed was

³VHULRXVOHDGHUVKLS´WRGRWKHZRUN%HLUDUHVLGHQWVIRXQGWKLVLQ0D\RU6LPDQJR158

Dondo

/LNHDOOFDVHV'RQGR¶V &0' EXGJHWgrew from 2004 to 2007, but it increased the most (78 percent on average) of all cases, with no dominant fee and tax in relation to overall receipts. Like Beira

DQG0DSXWR'RQGR¶VPXQLFLSDOVHUYLFHVDUHUDWHGKLJKO\E\IRFXVJURXSVDQG0HUFKDQW$VVRFLDWLRQV

However, of the four cases, Dondo has the lowest level of earned income to overall receipts ratio (22 percent), with central government transfers and a one-year transfer of aid shoring up the CMD. I would expect service to be low as a result because the accountability between fee payers and the municipality

ZRXOGEHZHDNGXHWR'RQGR¶VUHOLDQFHRQXQHDUQHGLQFRPH²namely, transfers²and absence of a tax lobby to pressure the CMD. However, Dondo, like Maputo and Beira, had strong mayoral leadership, and like Maputo, Dondo had targeted donor technical assistance that funded their participatory budget process. Mayoral leadership and targeted donor assistance helped the CMD make transparent resource allocation decisions for both earned and unearned income streams through the participatory budget process. This process provided the accountability pressure necessary to improve municipal service in the area of trash collection and market infrastructure.

157 Author interview, Mayor Daviz Simango, April 4, 2009, Beira; Focus group findings, Palmeiras I, March 17, 2009, Beira.

158 Focus group findings, Palmeiras I, March 17, 2009, Beira.

168

Dondo received almost USD$1 million for the life of the PDM project and USD$699 thousand through PADM. While both projects provide funding for infrastructure improvements, PADM phase I funds, awarded in 1999, provided the critical funding Dondo needed to undertake a participatory budgeting process that became the foundation for the C0'¶VLQIUDVWUXFWXUHLQYHVWPHQWVWUDWHJ\6LQFH

1996, prior to decentralization, Dondo had been pursuing funding for a participatory process to help it decide how to allocate funds for development within the then Conselho Executivo (Making Planning

Work 2009). In 1998, with donor assistance, the CMD surveyed households in all bairros to determine

ZKDWWKHFRPPXQLW\IHOWZHUHLQYHVWPHQWSULRULWLHVLQWKHLUQHLJKERUKRRGVIURPZKLFKVSUDQJ'RQGR¶V infrastructure investment plan (Roque and Tengler 2000). This plan allocates transfers and aid through

PADM and PDM. Both PDM and PADM funds and central government transfers provided support for

LQIUDVWUXFWXUHLPSURYHPHQWVWKDWUHVXOWHGLQLQFUHDVHGIHHFROOHFWLRQV7KURXJK3'0¶V0XQLFLSDO

Grants Program, Dondo received USD$954 thousand of which approximately half was for equipment used for trash collection. In 2006, the CMD received an additional USD$330 thousand infusion of

PDM and PADM funds that paid for office equipment, market infrastructure in markets like Samora

Machel and Mafarinha, urban planning in neighborhoods like Mafarinha, construction of bathrooms, public lighting, water foundations, and road construction. In addition, the CMD received a large infusion of state transfers for capital investments, likely road construction, in 2007 totaling USD $342 thousand.

A critical component of this performance improvement was mayoral leadership. Mayor

Manuel Cambezo has been mayor since 1998. His three town councilors for Finance, Urban Services and Environmental Management and Local Economic Development have been working for the CMD since 1998 either as town councilor or in a related position. This experience provides continuity to

'RQGR¶VSXEOLFDGPLQLVWUDWLRQ'HVSLWHDODFNRIRSHUDWLRQDOFRPSXWHUVWKHFinance town councilor is

DEOHWRGLVFXVVLQOXFLGGHWDLOJURZWKLQ'RQGR¶VRZQ-source revenues dating back to decentralization and the efforts that he has undertaken to increase and track fee collections. All three town councilors emphasized the link between improved services and increased fee collection. The Urban Services and

169

Environmental Management town councilor was approached by Cambezo just after he won the first mayoral election in 1998 to assume his current position. The Local Economic Development town councilor works his own farm, a fact especially pertinent since, in addition to overseeing the markets, his area of responsibility includes rural development and nutrition projects for those with HIV/AIDS.

When interviewed, Merchant Associations iQ'RQGR¶VWRSWZRPDUNHWVLQGLFDWHGWKH\ZHUH

VDWLVILHGZLWKWKH&0'¶VLQYHVWPHQWLQWKHLUUHVSHFWLYHPDUNHWVIURPWR$VWKHUHYHQXH picture suggested, infrastructure investment in markets through PDM and PADM outpaced investment in trash collection for good reason. The highest two markets account for the majority of fee revenue so the CMD can focus on improving services to a limited vendor population in contrast to trash collection

VHUYLFHVIRUDOOPXQLFLSDOUHVLGHQWVGLVSHUVHGWKURXJK'RQGR¶V several neighborhoods. Market fees constitute on average 25 percent of own-source revenue versus trash fees, which constitute only 7 percent of own-VRXUFHUHYHQXHDQGVRWKH&0'LVZLVHWRIRFXVRQYHQGRUV)XUWKHUPRUH'RQGR¶V

Finance town councilor indicated that his main responsibility was to organize receipts, mainly in the markets and that, since he began working at the CMD in 1999, Dondo has been successful.159 'RQGR¶V budgets provided evidence of this success. So, despite the smallest percentage of overall own-source revenue to receipts of the cases, the CMD provides a high level of municipal service. Donor assistance to develop a participatory budget process and the consistent municipal leadership to carry it out were the unique factors that drive performance in Dondo.

Matola

/LNHDOOFDVHV0DWROD¶V &00D EXGJHWgrew from 2004 to 2007. Furthermore, Matola had the fastest growth in earned income of all four cases, signaling that fee payers, in theory, were satisfied with the services the CMMa was providing. Fees constituted constitute the majority of own-source revenue (77 percent), with trash fees comprising the largest share. These revenue patterns suggest that

159 Author interview, Finance Town Councilor Anselmo Figuerieda, November 7, 2008, Dondo.

170

0DWROD¶VVHUYLFHVHVSHFLDOO\WUDVKFROOHFWLRQZLOOEHKLJKEHFDXVHWKH&00DUelies so much on these fees. However, focus groups and interviews indicate that corruption in both market management and trash collection, low municipal staff capacity, and lack of mayoral leadership resulted in poor service. I found that, unlike the other three cases, Matola did not have the same level of revenue diversification and coordination. Specifically, the majority of revenue came from central government transfers with little mayoral oversight or donor technical assistance to galvanize resources around improving trash collection and market management. Two donors identified the absence of mayoral leadership and municipal capacity as impediments to successful aid intervention, which might explain the low level of donor assistance in Matola. In the first example, one donor working in a neighboring municipality

LQGLFDWHGWKDWDODFNRIPXQLFLSDOFDSDFLW\GLVVXDGHGWKHPIURPZRUNLQJWKHUH7KH:RUOG%DQN¶V

PDM project similarly identified lack of municipal capacity and mayoral leadership as constraints to success of PDM. 2QHRIWKHOHVVRQVOHDUQHGIURP3'0ZDVWKDWD³VWURQJRZQHUVKLSIURPPXQLFLSDO

DGPLQLVWUDWLRQV´FRXOGKDYHLPSURYHGSURMHFWRXWFRPHV/DFNRIPD\RUDOOHDGHUVKLSLQWKH administration of the Municipal Grant Fund, specifically, was a challengH³3RRUFRPPLWPHQWRIWKH

Mayors in monitoring and coordinating the implementation of investments financed by the Municipal

*UDQW)XQG´ZDVFLWHGLQWKHILQDODQDO\VLV160 7KHUHSRUWFRQFOXGHGWKDW³the lessons from the (PDM) and other projects show that political leadership and commitment from the mayors/presidents is a precondition for a successful project design and management since the mayor has much greater ownership

DQGGLUHFWUHVSRQVLELOLW\DQGDFFRXQWDELOLW\´ ,PSOHPHQWDWLRQ&RPSOHWLRQ While the report itself did not single out Matola as having weak capacity, the CMMa was one of the municipalities that received an initial infusion of funds and the only municipality that did not receive a second tranche of

PDM funding.

160 6HH:RUOG%DQN¶V³,PSOHPHQWDWLRQ&RPSOHWLRQDQG5HVXOWV5HSRUW ,'$-35490 IDA 3549A) on a Credit in the Amount of US$33.60 million (SDR 26.70 million credit) to Republic of Mozambique for WKH0XQLFLSDO'HYHORSPHQW3URMHFW´5HSRUW1R,&5$)7:DWHUDQG8UEDQ6RXWKHUQ Africa Country Cluster 2, AFRICA, August 27, 2007, p. 35.

171

:KLOH'RQGR¶VGHSHQGHQFHRQWUDQVIHUVZDVPXFKKLJKHUWKDQ0DWROD'RQGRKDGDPD\RU- sponsored, donor-funded participatory budget process that coordinated revenue. Matola did not.

'RQGR¶VSURFHVVOLNH'U&RPLFKHZLWK3520$3872LQ0DSXWRLQVWLWXWHGDFFountability to residents and service improvements through donor technical assistance. While Beira did not have the same level of donor coordination, political accountability was high in Beira given its status as an opposition city. These politics forced Mayor Simango to live up to campaign promises as the first opposition mayor of Beira (see Chapter 6 for more details).

Focus group and Merchant Association interviews suggest that CMMa ranks low on municipal performance because of corruption and mismanagement. Although the CMMa has brought on qualified staff, corruption is a problem in vendor services, which I attribute to a combination of lack of capacity, adequate oversight, and political will. The president of the National Association of Municipalities of

Mozambique admitted that market fee collections often have two types of tickets²the official ones and non-official ones. In the morning, they sell the official tickets from which the fees are provided to the

CMMa and in the afternoon, they sell the unofficial tickets from which the proceeds benefit the fee collector.161 CMMa municipal staff confirmed that this is the practice and that market chiefs in some

PDUNHWVFKDUJHYHQGRUVVHOHFWLYHO\2QHGRQRUFDOOHGLW³WRWDOFRQIXVLRQ´LQWKHPDUNHWVDQGVWDWHGWKDW there was not the political will to make a change. The reporting structure makes supervising fee collectors difficult because they technically report both to the Finance and Patrimony and to the

Department of Markets and Fairs (DMF) so when planning staff call to verify figures, fee collectors would say they were in department or another to avoid being tracked down. The DMF town councilor has been working with the municipality since 1998 and has been the DMF town councilor for two mandates. However, since 2003, she has also assumed responsibility for Rural Development and Water so she is overburdened. She acknowledged that the CMMa does not have the resources in staff and transportation to oversee and collect fees. In 1998, when the CMMa started charging fees, the vendors held a strike. However, the only change that occurred since then is a modification of the regulations in

161 Author interview, President of the National Association of the Municipalities of Mozambique, March 2, 2009, ANAMM office, Matola.

172

2001 to increase vendor fees by 2 MTn Although the CMMa had taken recent steps to hire highly qualified staff which include a Ph.D. economist, comparatively, municipal staff capacity remains weak.

Mismanagement and lack of political will extends to trash collection. Like all cases, the

CMMa has to contend with old equipment. Municipal staff reported that they have eight tractors, a bulldozer and skiploader but they often breakdown, which stymied collection.162 When trash fees were introduced, there were sporadic protests but the municipality did nothing to alter its approach in charging flat fees or work with donors to develop a technical solution, like scaled fees, as did Maputo.

Municipal staff similarly indicated that, although the CMMa faces substantial challenges to collecting trash, that the problem was lack of political will. A donor who had been working in the area of solid waste management indicated that the level of public administration and professionalism were so low that they pulled out.

The case of Matola conformed to one aspect of fiscal theory: it had majority shares of unearned income. The CMMa received on average the largest share of central government transfers in comparison to the other three municipalities. However, own-VRXUFHUHYHQXHJUHZGXHERWKWR³QRQ-VSHFLILF´IHHV which are difficult to link back to any specific group of fee-payers and trash fees. The evidence suggested that the CMMa did provide services for which it effectively collected fees. However, the lack of revenue diversification and coordination created an environment of low accountability and low performance in the two test areas. In the absence of strong tax lobbies and the capacity to provide services, unearned income in the form of aid played a determinant role in generating own-source revenue across the board as evidenced by the World Bank and GTZ projects I discussed. However in

Matola, the combination of a lack of political will to change corrupt practices and mismanagement, high dependence on central government transfers, and low level of aid as a share of overall municipal

UHYHQXHVDIIHFWHGWKH&00D¶VOHYHORISHUIRUPDQFHVXFKWKDWWKHRZQ-source revenue generated was not used to improve performance. Trash service was selectively provided. Fee collection practices in

162 ,QWHUYLHZ7HFKQLFLDQLQWKH6DQLWDWLRQDQG6DOXEULW\7RZQ&RXQFLORU¶V2IILFH1RYHPEHU Matola.

173

the markets²specifically, double charging vendors and poor municipal oversight over market chiefs² mirror the low-level corruption reported in trash collection. All these factors indicate that growth in earned income alone is not a sufficient determinant of accountable, high-performing government when governments are still reliant on high levels of government transfers and have weak mayoral leadership.

Conclusion

The purpose of this chapter was to examine the revenue pressure exerted by earned and unearned income over municipal performance. Specifically, I analyzed how the incentives created by flows of own-source revenue, central government transfers, and aid affected municipal performance and accountability to fee payers. I hypothesized that municipalities with majority shares of earned income will have high performance in the test areas. This was hypothesized to work as follows: governments are thought to respond to taxpayers that contribute a dominant share of state revenue because these taxpayers can effectively bargain with the state, even if they do not use this pressure, because the state relies on this revenue, creating incentives for better performance. Overall, I observed two dynamics at play here. First, decentralization is working. It provided the authorities to keep and collect more local revenue. It also is created incentives for local ownership of service provision, evidenced by the various town councilors in Beira, Dondo and Maputo emphasizing the link between the need to increase fees on order to improve service delivery. Second, while these cases revenue profiles consistently high shares of central government transfers and aid, own-source revenue increased. Specifically, I saw three encouraging trends.

First, I found municipal performance improved absent a tax lobby to pressure it. Specifically, I found that a variety of fees and taxes constituted own-source revenue across all cases. While trash and market fees constituted the largest component of own-source revenue, it was on average one-third of own-source revenue. This is an important finding because it suggests that performance improves in the absence of dominant lobby groups to pressure government performance improvement.

174

Second, I hypothesized that large and growing shares of earned income would be associated

ZLWKEHWWHUPXQLFLSDOVHUYLFHVEHFDXVHWKHPXQLFLSDOLW\ZRXOG³H[FKDQJH´EHWWHUVHUYLFHVIRUIHHVDQG would be more responsive to fee-payers because the municipality would depend on them for revenue.

However, I found that large (and growing) shares of own-source revenue was not a sufficient condition to improve municipal service delivery, as evidenced in Matola, nor was it a pre-requisite for strong performance, as evidenced in Dondo. While DonGR¶VUHYHQXHSURILOHKDGWKHORZHVWRZQ-source revenue to overall receipts (22 percent), highest aid receipt (25 percent) and 50 percent in central government transfers, trash collection and vendor services were still considered to be high. Matola, on the other hand, relied predominately on central government transfers, with the highest rate of transfers to overall receipts (63 percent). My evidence indicated that high municipal performance in the three cases (Maputo, Dondo, and Beira) was due to combination of specific donor technical assistance projects and mayoral leadership, which guided the investment of all sources of revenue.

My last hypothesis was that governments that rely primarily on unearned income²specifically central government transfers and aid²would have little incentive to provide services so I expected government performance in these cases to be low. I found this to be true in Matola but not in Dondo, as stated above. I found unearned income²specifically, targeted donor assistance that helped coordinate aid, transfers and own-source revenue²improved municipal performance in municipal cases with strong mayoral leadership. While aid as unearned income can disincentivize government performance because of the absence of the pressure fee payers in theory exert, donors require accountability. This effect was evidenced in Maputo and Dondo²both of which had overarching donor projects that framed programming and how all funds were spent.

I also found that leadership mattered. Dondo, Beira, and Maputo had strong mayors with technical backgrounds who were able to marshal resources to improve municipal service in the two test areas. In Maputo, Dr. Comiche put together a team that was worked closely with the GTZ to revise the trash fee schedule and institute a sustainable plan to collect scaled fees and phase donor assistance out in

WKHORQJUXQ3520$3872ZKLFKLV0DSXWR¶VOong-term community and economic development

175

strategy funded through the World Bank, is a cornerstone to municipal service improvement. Dr.

&RPLFKH¶VUROHLQJXLGLQJWKLVSURFHVVZLOOEHGLVFXVVHGLQWKHQH[WFKDSWHU0D\RU&DPEH]RLQ'RQGR was head of BeLUD¶VPXQLFLSDOJRYHUQPHQWEHIRUHGHFHQWUDOL]DWLRQSXWWRJHWKHUDWHDPDQGZRUNHG closely with the Austrian Development Cooperation to institute the participatory budgeting process still in use today. Lastly, Mayor Simango in Beira was able to immediately mobilize own-source revenue and staff resources to collect trash in his first few months in office.

The analysis of this chapter touched on the role politics played in crafting the opportunities mayors had to institute change. Beira, an opposition municipality, received the smallest share of central government transfers to overall receipts of all four cases (42 percent on average) from 2004 to 2007.

This implies that Beira is the least dependent on central government transfers of all four cases, which

PD\EHWKHUHVXOWRI%HLUD¶VSROLWLFDOHFRQRP\'RQGR¶VKLJKVKDUHRIDLGDQGWUDQVIHUVDQGORZVKDUHRI

HDUQHGLQFRPHPD\EHLQIOXHQFHGE\'RQGR¶VSUR[LPLW\WR%HLUDDQGWKHLQFHQWLYHWKH)5(/,02 government may have to make this FRELIMO municipality successful. Maputo, in addition to benefiting from its position as the capital and proximity to a major economy in South Africa, may also be subject to the same political pressure by the FRELIMO party to be a national model because it is home to donors and diplomats active in Mozambique. The siphoning of market fees, which arguably resulted in Matola having the lowest share of market fees to overall receipt, may be because of the lack of political competition and the focus on neighboring Maputo. The following chapter will explore these political dynamics to understand how politics affects municipal performance in these four cases.

176

Table 5.5: Expected v. Actual Outcomes

Hypothesis: Decentralization creates incentives that motivate municipal government officials to act and opportunities for residents to pressure them.

RESOURCE PRESSURE

Revenue:

177

Table 5.6: Across Case Averages: 2004-2007

Average, Budget Revenue Market Trash Central Aid 2004-2007 autonomy fees to fees to gover nment dependence* own- own- dependence (Aid to overall (Own- sour ce sour ce receipts) sour ce t o receipts receipts (Transfers to overall (fees)+ (fees)+ overall receipts) receipts) Maputo $8.9 57% 10% 17% 44% 9% million (27%) (41%) (+34%) Beira $5.2 49% 19% 12% 42% 9% million (25%) (16%) USD (+63%) Dondo $755 22% 23% 7% 50% 25% thousand (27%) (8%) USD (+78%) Matola $1.9 35% 9% 29% 63% 9% million (12%) (38%) USD (+70%) Sources: CMM Contas de Gerência de 2004, 2005, 2006, and 2007; CMB Contas de Gerência de 2004, 2005, 2006, and 2007; CMD Contas de Gerência de 2004, 2005, 2006, and 2007; CMMa Contas de Gerência de 2004, 2005, 2006, and 2007. Exchange rate used USD$1: 26.5MTn.

*Aid to overall receipts column presents findings from both analysis of municipal budgets.

+Itemized trash and vendor fee data for 2004 unavailable for Dondo so average is from 2005-2007.

178

Table 5.7: Selected aid projects, 2004 to 2007

AID Selected aid Description Donor Disbursements Aid (USD) investment autonomy Maputo M unicipal Municipal grant fund World Bank $14 million Low Development for consulting and (2001-2005, $1 (Ministry of Program (PDM ) infrastructure services million in 2004- Tourism) (markets, drainage, 2005) sewage, sanitation and PROM APUTO roads). World Bank $3.7 million Build institutional (2007, $43 High AGRESU capacity and million Phase 1) (Maputo) infrastructure GTZ $999,627 (2007)

Establish and High implement a modern, (Maputo) integrated and sustainable waste management system Beira Investment in Improve the EC $1.4 million High (Beira) three conditions of families (2004-2007) neighborhoods in Goto, Munhava, and Muavi. M unicipal Development Municipal grant fund World Bank $2.5 million Low Program (PDM ) for consulting and (2001-2007) (Ministry of infrastructure services Tourism) (markets, drainage, PADDEL sewage, sanitation and Italian $718,770 (2007, roads). Cooperation Beira share) Low (MAE)

Strengthening public UNICEF service provision and socio-economic UNICEF $490,522 (2006- development. 2007, Beira share) High (CMB)

Increasing access to safe water and sanitation to five municipalities. Dondo PADM Support of Austrian $699,651 (2005- Low participatory processes Development 2007) (Provincial in 2 municipalities and Cooperation Directorate of 5 districts. Public Works and Housing) M unicipal Municipal grant fund Development for consulting and World Bank $954 thousand Program (PDM ) infrastructure services (2001-2007) Low (markets, drainage, (Ministry of sewage, sanitation and Tourism) roads). 179

Table 5.7 (Continued): Selected aid projects, 2004 to 2007

Matola Institutional Support for urban Spanish $2.2 million Low (MAE) Suppor t to planning, food security Cooperation (2004-2007, Mozambican and training of civil Matola share) M unicipalities servants in two Program municipalities and an institute

Municipal Legal and Institutional World Bank $24 thousand Low Development Reform (Matola share) (Ministry of Program (PDM)* Tourism)

Greater Maputo Environmental DANIDA $714 thousand Low (City of Area-M atola management strategy (2005-2007, Cuamba) and implementation estimated award for time period) 6RXUFHV2'$02=GDWDEDVHDXWKRU¶VFDOFXODWLRQVIRULQGLYLGXDOPXQLFLSDOVKDUHV H[FKDQJHUDWH applied USD$1:26.5 MTn). The top projects in Maputo, Beira and Dondo were large infrastructure LQYHVWPHQWVZKLFK,UHPRYHG:RUOG%DQN³,PSOHPHQWDWLRQ&RPSOHWLon and Results Report (IDA- 35490 IDA 3549A) on a Credit in the Amount of US$33.6 million to the Republic of Mozambique for WKH0XQLFLSDO'HYHORSPHQW3URMHFW´$XJXVW$)7:DWHUDQG8UEDQ6RXWKHUQ$IULFD Country Cluster 2, AFRICA, accessed at www.worldbank.org on July 22, 2010. Of the cases studied, only Maputo, Dondo and Beira were included in the Municipal Grant Fund. To demonstrate the level of aid autonomy in each municipality, I examined the three top-funded projects in each of the four case study municipalities and projects related to decentralization to ascertain two points: first, what was the nature of large-scale donor investment and second, which entity²a central government ministry or the municipality itself²was designated the channel of delivery. In Maputo, Beira and Dondo, the three top projects are large-scale infrastructure projects: water and sanitation in Maputo and Beira and railroad reconstruction and primary education in Dondo, which I excluded. In Matola, the top-funded project is a food security program through the ADB. Provincial or central government agencies are the designated counterparts for these projects, which is why the projects are designated low on aid autonomy. Bolded projects are included in the municipal budgets. Aid totals reflect what I found in both donor reports and in municipal budgets and may not correspond.

0DWRODZDVLQFOXGHGRQO\LQ&RPSRQHQW6XEFRPSRQHQW³,PSOHPHQWDWLRQRI3LORW([SHULHQFH´ Management of Human Resources, along with Maputo, Manhiça, Dondo, Beira, Nacala, Quelimane, Pemba, Nampula and Xai Xai. Matola was not included in the Municipal Grants Fund. I estimated 0DWROD¶VVKDUHE\HYHQO\FXWWLQJ86'PLOOLRQIRUWKH/HJDODQG,QVWLWXWLRQDO5HIRUP&RPSRQHQW by four subcomponents, five processes, two outputs and three cities that received assistance under the Management of Human Resources output.

180

Table 5.8: Conclusion

REVENUE M UNICIPAL PERFOM ANCE Expected Actual

M APUTO

Markets Majority Majority earned income, coordinated, targeted High Earned donor assistance through PROMAPUTO and Trash Income *7=¶V$*5(68VWURQJPD\RU Collection

BEIRA

Markets Majority Majority earned income and strong mayor High Earned Trash Income Collection

DONDO

Markets Majority Majority unearned income, coordinate donor High Earned assistance, highest levels of donor assistance, Trash Income strong mayor Collection

M ATOLA

Markets Majority Majority unearned income, no donor Low Unearned coordination, no mayoral leadership. Trash Income Collection

181

VI: VOTING EFFECTS ON SERVICE DELIVERY

Inkosa yinkosa ngabantu (A leader is only a leader because of his people).

Chi-Shona saying163

I have seen a change since 5 years, specifically in the mandate of this president (Eneas Comiche). But they (FRELIMO) cut him loose because of the politics. He was cut loose because of politics not because of technical capacity to get the job done. Everything has changed.

Interview with a Maputo Vendor, October 1, 2008.

Introduction

In this chapter, I examine how elections affected municipal performance. Specifically, I hypothesized that competitive elections incentivize mayors to improve municipal services to constituents (see Table 5.1). As the above-stated quotes suggest, I found that dominant party politics affected the ability for voters to exert political pressure (and express political preferences) through their votes in thereby muddying the incentive mayors have to improve municipal services to all voting constituents. In all cases, I found that technocratic mayors²specifically, mayors who were good public administrators with a background in management²improved performance across cases regardless of party. Lastly, mayoral performance a mattered most for garnering constituents votes for re-election in the opposition, not dominant, municipalities.

As I discussed in Chapter 2, the democracy literature argues that elections affect performance because elected officials compete for (re-)election, compelling these officials to adhere to voter demands for better, more responsive, more accountable government (Dahl 1971; Diamond 1999;

Lijphart 1999; Huntington 1991; Fukuyama 1992). I found that two questions surfaced in my study of

163 $VTXRWHGE\,UDr%DSWLVWD/XQGLQLQ³&LWL]HQ3ROLWLFDO3DUWLFLSation in a Changing Society: The &KDOOHQJHRI$EVWHQWLRQVLQWKHDQG(OHFWLRQVLQ0R]DPELTXH´$XJXVW

182

how elections affected performance in Mozambique that merited further exploration. First, in the literature, democracy performance is linked the quality of elections rather than the quality of governance. The literature underemphasizes factors--like leadership and technical capacity--that led to improvement in government performance beyond the incentives elections provide elected officials.

Elections incentivize elected officials to act in way that will enable them to be re-elected which, while including fulfilling campaign promises that improve government performance, also require them to act on behalf of their party and serve key constituents and lobbies. These political incentives may or may not improve overall municipal performance for all voters. I found that the municipal cases I studied that had high performance in the test areas had mayors with technical backgrounds who knew how to manage a team. In Beira, the mayor is a civil engineer and had worked on municipal contracts. In

Maputo, the mayor was a former executive with the Bank of Mozambique, had extensive management experience, and is a trained economist. In Dondo, the mayor was the former city administrator of Beira prior to decentralization. In all three cases, mayors put together teams, which helped them implement the improvements in municipal services on which they campaigned.

Second, while, in theory, political competition generated by multiple candidates and parties vying for constituent votes can lead to improved governance, the dynamic which creates these incentives to compete²namely, political competition in a multi-party state²is absent in

Mozambique.164 As such, to understand how elections incentivize mayors to improve municipal performance, I examined not only interparty competition but also intraparty competition. As I discussed in Chapter 1, I argued that the more competitive the election²meaning the more competitors and the greater the political space as measured by both mayoral and party turnover²the more re-election means to a mayor and the harder he will work to deliver on his campaign promises. I selected Beira, Maputo,

Dondo and Matola based on having variation on levels of competition to analyze how political pressure incentivized mayoral performance. Based on the levels of interparty and intraparty competition, I expected political pressure in Beira to be high since both the mayor and party switched in the 2003

164 While there are over 30 political parties in Mozambique, the same party²FRELIMO²has been in power for close to three decades, monopolizing national and local elected offices.

183

election and this turnover would have incentivized the incoming mayor to improve municipal performance. I found this to be the case in Beira. I expected political pressure to be medium in

Maputo since only the mayor, not party, switched in the 2003 election. This turnover would have incentivized the incoming mayor to improve municipal performance but I suspected the monopoly of the dominant party would have more influence over the mayor than voters so broad-based improvement to constituents would be secondary to servicing the party. I found that the medium rating was consistent with the evidence that the mayor sought to improve municipal services to his constituents at the expense in some instances of the party interests. The mayor did improve services to voters in the 2003 mandate

EXWLQFKDOOHQJLQJWKHGRPLQDQWSDUW\¶VDFFHVVWRUHVRXrces, he jeopardized his nomination and was subsequently not re-nominated in 2008. I expected political pressure in both Dondo and Matola to be low because there was no turnover in mayor or party. I found that, in Dondo, while political pressure resulting from elections was low, as a FRELIMO stronghold, political pressure from the party to provide municipal services was high for Dondo. Dondo was indeed successful in delivering satisfactory services in the two areas studied. The Matola case was excluded from review for this period because the

PD\RUHOHFWHGLQGLHGGXULQJWKHPDQGDWH ³)5(/,026WULYLQJ´; ³0RUUHX&DUORV7HPEH´

2007).165

Lastly, one curious parallel that surfaced during this study was how the mayors of two of

0R]DPELTXH¶VKLJK-performing municipalities were being ostracized from their respective parties in the lead up to the 2008 mayoral elections and the reaction of both parties and these mayors. Both mayors were nominated by their parties in 2003 and were considered to have fulfilled or be fulfilling campaign promises heading into the 2008 nominations process. However, the leadership of both parties did not re-nominate either. In Maputo, this meant that Mayor Comiche did not receive the FRELIMO nomination and in Beira, RENAMO retracted its nomination of Mayor Simango. In Maputo, Mayor

Comiche did not challenge this decision whereas in Beira, Mayor Simango did, ran as an independent,

165 Carlos Tembe died December 20, 2007 of hepatitis-related causes. He was first elected as mayor of Matola in 1998 and subsequently re-elected in 2003. The interim mayor, Maria Vincente, formerly the chairperson of the Municipal Assembly, ran for the FRELIMO party nomination and lost to the nomination to Arão Nhacale. Nhacale won the 2008 Matola mayoral election

184

won and created a national party. The democracy literature suggests that a strong opposition party motivates higher levels of performance by fostering competition (Manning 2005; Manning 2010;

Schedler, Diamond and Plattner 1999; Przeworski, Stokes, Crook and Manor 1998). However, I found that weakness and disorganization of the opposition party actually permitted greater political competition and higher municipal performance because the Beira mayor relied, not on the opposition party leadership, but on the voters.

I begin this chapter by analyzing the 2003 mayoral elections. Then, I examined how mayors during the subsequent mandate (2003-2008), made changes to municipal administration that improved

WKHPXQLFLSDOLW\¶VSHUIRUPDQFHRIWUDVKFROOHFWLRQDQGYHQGRUVHUYLFHVXVLQJLQWHUYLHZVQHZVSDSHU accounts, and focus group findings. From there, I compared performance to the results of the 2008 mayoral nominations and subsequent elections to see if performance played a decisive role in the (re-) election of mayors in my four cases. I conclude by summarizing these findings and reiterating how elections affected municipal performance in these three cases.

2003 Mayoral Elections

Since the first multiparty elections were held in 1994, Mozambican political parties have proliferated to roughly 30 parties (Carbone 2005,422). However, FRELIMO remains the dominant party with RENAMO remaining the distant second and progressively weaker opposition party.

Although the first multiparty municipal elections were scheduled for 1997, they did not occur until

1998, after being postponed three times due to RENAMO-FRELIMO party politics, resulting in only a

15 percent voter turn-out (see Table 5.3) (Cuereneia 2001).

RENAMO abstained from the 1998 municipal election making the 2003 municipal elections the first cycle based on the performance of a previously-elected mayor and first election in which the two major parties actually competed. Although FRELIMO won all the mayoral races in the 1998 elections, by 2003, the FRELIMO leadership jettisoned 20 mayors in favor of a clean slate, re- nominating only 13 incumbents to run again based on performance (Nuvunga 2004). Experience with

185

municipal administration, however, was not the dominant criteria for filtering FRELIMO candidates with only one-fifth of the nominees possessing any connection to municipal government in the preceding five years (Ibid). In 2003, FRELIMO won 28 mayoral positions and dominance in 29

Municipal Assemblies with RENAMO capturing five mayoral positions and majorities in four

Municipal Assemblies.166 Successful incumbents ran on performance whereas successful challengers ran anti-corruption campaigns against the incumbents. In the three cases studied, voter turnout was low, hovering between 20 to 30 percent (see Table 5.3). As such, this was the appropriate baseline from which to launch an investigation of how political accountability affected municipal performance between the 2003 and 2008 elections by pressuring mayors to fulfill campaign promises.

Maputo

Political pressure to increase municipal services²such as trash collection and conditions in informal markets²in the wake of the 2000 floods was a driving force in the to the 2003 elections.

Entering the 2003 elections, the FRELIMO incumbent Mayor Cananá was nominated in 1997 to be the president of the Executive Council before decentralization required elections of mayors (Noticias

2006).167 He ran and won the first mayoral elections in 1999. However, his Administration was

PDUNHGZLWKFRUUXSWLRQZKLFKVWHPPHGIURP&DQDQi¶VLQH[SHULHQFHZLWKUXQQLQJWKHFDSLWDODIWHU having been mayor of the smaller, inland city of Chimoio.168 Cananá attrLEXWHG³DEVROXWHSRYHUW\´DV

WKHFKLHIFDXVHRIWKH³PHQGDFLW\DQGFULPLQDOLW\´WKDWSODJXHGWKHFDSLWDOFLW\GXULQJKLVPDQGDWH

³0HQGLFLGDGHH&ULPH6mR3UHRFXSDQWHV´ 7KH&DQDQi$GPLQLVWUDWLRQZDVXQDEOHWRGHDOZLWK

166 Mozambican Constitutional Council, certified results for 11/19/2008 and 11/19/2003 municipal elections. As Primeiras Autarquias em Mocambique: Realidade, Conquistas, Constrangimentos, Perspectiva: MAE and AWEPA, 2003 and Adriano Nuvunga, Cartografia do Poder Local 2003-2008 (AWEPA, Maputo, 2004 funded by Reina de Suica) for the first elections held 6/30/1998.

167 Cananá was a member of the Political Commission and Central Committee of the FRELIMO party and was the former governor of Manica province.

168 (FRQRPLVWDQG8(03URIHVVRU/LQR0DUTXHVDVTXRWHGLQ³,OXVmRGHVXFHVVR´H[FHUSWHGIURP Savana, December 14, 2007, 3. Accessed 10/29/10 from macua.blogs.com.

186

road maintenance, modernization of the markets and trash collection, which led to corruption scandals, resulting in the nomination of Dr. Comiche as the FRELIMO candidate (Wilson and António 2001,44).

Trash collection and organizing informal vendors became two important issues of the 2003 election. Dr. Eneas Comiche came from the private and academic sectors with a reputation as a technocrat. Running on a reformist agenda, Dr. Comiche took on trash collection and organizing

LQIRUPDOPDUNHWVDWWDFNLQJFRUUXSWLRQKHDGRQ³,I,DVNIRUDOLFHQVH«,KDYHWKHULJKWWREHDWWHQGHG

WRRQWLPH´&RPLFKHUHODWHGLQDQLQWHUYLHZ³DQGQRWZKHQWKHPXQLFLSDOLW\IHHOVOLNHDWWHQGLQJWRPH

DQGQRWRQO\ZKHQ,EULEHDPXQLFLSDORIILFLDO´ ³,QGHSHQGHQW&DPSDLJQV´, 11). Mayor

&RPLFKH¶VElectoral Manifest outlined several initiatives, which included working with community organizations on trash collection capacity and working with the Maputo residents on increased fee collection. Furthermore, he highlighted working with informal vendor associations to increase sanitation

FRQGLWLRQVDQGPDUNHWRUJDQL]DWLRQ³,KDYHKHDUGLWVDLGWKDWPRQH\LVFROOHFWHGEXWWKDWQRERG\NQRZV where it goes. I have heard that more than enough money is collected in market fees to ensure that the toilets work, and that there is water available, to keep the markets clean, and to guarantee security, yet

QRQHRIWKLVLVGRQH´&RPLFKHGLVFXVVHG ³&RPLFKHSURPLVHVPXQLFLSDORPEXGVPHQ´, 11).

Comiche ran a distinctly different campaign in which his manifesto reDGOLNHD³YHLOHGFULWLFLVPRIKLV

SUHGHFHVVRU´SURPLVLQJWRDSSRLQWD³WHDPRIFRPSHWHQWKRQHVWDQGLQFRUUXSWLEOHSHRSOH´ ³)5(/,02

3URPLVHV%DWWOH$JDLQVW&RUUXSWLRQ´, 10; ³,QGHSHQGHQW&DPSDLJQV$OPRVW,QYLVLEOH´, 11).

He won the general election with 75 percent of the vote (see Table 5.4).

Beira

Like the 2003 mayoral nomination in Maputo, FRELIMO did not support the outgoing mayor,

&KLYDYLFH0XFKDQJDJHZKRKDGDOVREHHQDFFXVHGRIFRUUXSWLRQLQRIILFH ³)5(/,02µFDSLWXOD¶ ´

2000; ³0R]DPELTXH6LPDQJR3XEOLVKHV'RFXPHQWV´; Mosse 2006, 11).169 However, he began

169 Both Muchangage and Lourenço Bulha were accused of using their authority as Mayor and President of the Municipal Assembly during the 1998 mandate to inappropriately use municipal property. In one instance, the press reported that Muchangage transferred municipal property ownership to Bulha; in another, Muchangage was involved with the transfer of rent collection on municipal property to the 187

informally campaigning and faced competition for the party nomination by former Municipal Assembly

President, Lourenço Bulha. The competition turned acrimonious resulting in what one political analyst

FDOOHGDQ³XQSUHFHGHQWHGLQWHUQDOZDU´ 1XYXQJD &RQVHTXHQWO\WKH)5(/,02SDUW\QRPLQDWHG a third candidate, Djalma Lourenço, who eventually came under a cloud of suspicion because he was fired in 1991 from a civil servicHSRVLWLRQIRUKDYLQJDOOHJHGO\PLVDSSURSULDWHGVWDWHSURSHUW\ ³'LG

)5(/,02FDQGLGDWHVWHDOVWDWHSURSHUW\"´,12-13).

Daviz Simango, a member of a smaller party, which was part of the RENAMO-UE coalition, was asked to run and he stipulated that, if he were to accept the RENAMO nomination, he was going to be a serious candidate (Sitoe, Matsimbe, and Pereira 2005,19).170 When approached in 2003 to be the

5(1$02FDQGLGDWHKHVWDWHGWKDW³,ZDVJRLQJWRZRUNGLIIHUHQWO\«,ZDVQRWJRLQJWRZRUNZLWKOD]\

SHRSOHZKLFK,NQHZZDVJRLQJWRFDXVHDVRUWRIVKRFN´ 0R\DQD,21).171 Like Maputo, the

GRPLQDQWLVVXHLQ%HLUDZDVWUDVKFROOHFWLRQDQGWKHFRUHFDPSDLJQLVVXHZDVFRUUXSWLRQ³0HPEHUVRI

WKH)5(/,02SDUW\«ZHUHWRRLQYROYHGLQSROLWLFVDQG WKHUH ZDVQRVSLULWRIZRUN´6LPDQJR recalls.172 6LPDQJRZRQWKHHOHFWLRQE\SHUFHQWRIWKHYRWHDJDLQVW'MDOPD/RXUHQoR¶V percent (see Table 5. ³)5(/,02&KRRVHV,WV6RIDOD&DQGLGDWHV´; Ruigrok 2005).

Dondo

Manuel Cambezo, one of the thirteen mayors FRELIMO re-nominated in 2003, brought to the job several years of experience in neighboring Beira, as the city administrator of the Conselho

Executivo, WKHFHQWUDOL]HGDGPLQLVWUDWLYHSUHGHFHVVRUWRWRGD\¶VSROLWLFDOO\GHFHQWUDOL]HGConselho

former Beira FRELIMO secretary between 1999 and 2003. Muchangage was later appointed a District Administrator of Angónia, Tete Province.

170 Simango was a member of the PCN (National Congress Party) of which his brother was the head. National Congress Party or PCM (Partido Congresso Nacional), headed by his brother.

171 8ULD6LPDQJR'DYL]6LPDQJR¶VIDWKHUZDVRQHRIWKHIRXQGHUVRI)5(/,02SUHVHQWLQ'DU(V- Saalam on June 25, 1962 and later was deputy President from 1962 to 1969. He and his wife were GHFODUHG³UHDFWLRQDULHV´RIWKH5HYROXWLRQDQGH[HFXWHGin 1977 at the Majune re-education camp in Niassa province. Author interview, Daviz Simango, Mayor of Beira, April 4, 2009, Beira.

172 Author interview, Daviz Simango, April 4, 2009, Beira.

188

Municipal, in the 1980s. In 1998, Cambezo ran unopposed and won in a municipality with a high level of voter turn out (53.63 percent) in comparison to national levels, which were extremely low on average

(Hanlon 1998). 173 Political analysts speculated that ballot box stuffing occurred in Dondo because

Dondo was the only municipality in which the turnout exceeded 50 percent (Ibid). 174

In the subsequent 2003 election, Cambezo ran against Joaquim Greva (RENAMO) and Matias

Traquino (IPADE) (see Table 5.3). Dondo was already regarded as a high-performing municipality

LQWHUQDWLRQDOO\E\GRQRUVVHOHFWHGDVWKH³EHVWSHUIRUPLQJPXQLFLSDOLW\LQ´LQSDUWEHFDXVHRIWKH

$XVWULDQ'HYHORSPHQW&RRSHUDWLRQ¶VLQYHVWPHQWLQ'RQGR¶VSDUWLFLSDWRU\EXGJHWLQJSURFHVV

(Manndorff, Nöst et al 2004, :LWK&DPEH]R¶VH[SHULHQFHKHZDVDEOHWRHDVLO\GRPLQDWHWKH

Municipal Assembly, which lacked the technical expertise to run programs and analyze budgets

(Nuvunga, Mosse, Varela 2007,15-16). Dondo had high abstention rates in this election, with 38,635 registered voters and only 11,417 casting their votes or a 29.55 percent turnout, suggesting a low level of competition in this election (Ibid,15-16; ³0XQLFLSDOHOHFWLRQV´,21). Manual Cambezo won

61.61 percent of the vote (6,641 votes) (see Table 5.3).

In conclusion, all three mayoral candidates ran and won on platforms to improve municipal service. In Maputo and Beira, trash collection and improved market infrastructure was a cornerstone of candidate platforms. In DoQGRWKHPD\RUFRXOGWDNHVRPHFUHGLWIRUWKHPXQLFLSDOLW\¶VSHUIRUPDQFH

While voter turn out for the three cases was low, the 2003 municipal elections were the first multiparty elections in which the major opposition party participated. The prospect of 5(1$02¶VSDUWLFLSDWLRQ may have motivated FRELIMO to jettison political liabilities and focus on candidates that could deliver on campaign promises to improve municipal services. This is arguably an example of how elections incentivized the dominant political party to back mayors that could improve municipal services.

173 The national average voter turnout was 14.58 percent or 286,659 out of 1,965,530 in 1998.

174 Some speculate that poll workers in favor of Cambezo assumed that he could win only if more he UHFHLYHGPRUHWKDQSHUFHQWRIWKHYRWH*XLGRYDQ+HFNHQ³%DOORW%R[6WXIILQJLQ'RQGR"´

189

Performance from 2003 to 2008

Maputo

The mandate of Mayor Comiche demonstrated proficiency with technical details and management most evident in his leadership style and pursuit of a unified donor investment and economic development policy. Starting with his campaign, he presented his Electoral Manifest in all of the bairros, incorporating community feedback. This style of holding hearings carried into his mandate, which became known DVWKH³presidência aberta´ RSHQPD\RUDOPDQGDWH DQG³governação aberta´ RSHQJRYHUQDQFH LQZKLFKFLWL]HQVZHUHLQYLWHGWRSDUWLFLSDWH+HKDGRSHQIRUXPVHYHU\WR

3 months, which were advertised in the radio, TV and print press and were coordinated through his town councilors.

As Mayor, Dr. Comiche had the right under law to appoint half of his town councilors from any sector he chose. He selected a group of town councilors based on expertise. 175 His executive core was comprised of engineers, accountants, doctors, linguists, and businessmen who had common ideas and a diversity of experience. They met every day at 8 am for no more than an hour at a meeting chaired by the Mayor and, in this way, became a cohesive team. Starting with this basic core, Dr.

Comiche began reforming the way City Hall was run. Mayor Comiche cut municipal jobs from 3,112 in 2005 to 2,662 by 2008, a 14 percent reduction. 176 He strongly encouraged his town counselors to follow suit and introduced performance exams IRUPXQLFLSDOMREV%\WKHHQGRI0D\RU&RPLFKH¶V

175 For three powerful posts, he selected a doctor, a business owner and a Finance ministry official who was detailed to the CMM, for the posts of Town Councilor of Health and Salubrity, Town Councilor of Commerce and Industry, and Town Councilor of Finance. The Town Councilor for Health and Salubrity is a doctor and was teaching at the University Eduardo Mondlane Faculty of Medicine for 10 years whereas the Town Councillor of Commerce and Industry was a business man for 30 years. All three were experienced men in their fields who came into MaSXWR¶V&LW\+DOOXQSUHSDUHGWRXQFRYHU there was no infrastructure to even start work. Dr. Schwalbach, the Town Councilor for Health and Salubrity, named that one of his primary challenges when he began his term in February 2004 was building an office infUDVWUXFWXUHIURPVFUDWFK+HVKDUHG³ZKHQ,FDPHLQ,KDGRQHVHFUHWDU\DQGD FKDLU7KDWZDVLW1RWHOHSKRQHQRFRPSXWHUQRDUFKLYHVQRQRWKLQJ6RPHGLGQ¶WHYHQKDYHDQ RIILFH´ Author interview, Town Councilor for Sanitation and Salubrity, Maputo City Hall, October 7, 2008.

176 Author interview, Dr. Eneas Comiche, Maputo, December 12, 2009.

190

mandate, City Hall was wired and offices were well-marked and well-staffed by professionals who responded to public inquiries in a methodical fashion. Lastly, the Comiche Administration proposed a new public participation process for budget approval to commence in 2009, which the Municipal

Assembly subsequently approved.

7KHVHFRQGVLJQLILFDQWSLOODURI&RPLFKH¶VDGPLQLVWUDWLRQWKDWDIIHFWHGERWKPDUNHW management and trash collection was his work with donors²specifically, the World Bank through the

PROMAPUTO project and the GTZ through AGRESU. PROMAPUTO, a unified economic development strategy, which addressed the main concerns of his Electoral Manifest, was the result of three large consultations held from 2004 to 2008 of participants who were members of civil society, the private sector, academia and members of the public. The purpose of the consultations was to create a

10-year, multi-sector development strategy for Maputo. This was significant in scope because there were no planning documents of this kind in existence. 177 Beginning with a planning meeting in the fall of 2005, over 100 participants from the private sector, embassy and donor community, civil society and the Maputo (CMM) gathered to discXVV0DSXWR¶VPDMRUGHYHORSPHQWQHHGVRYHUWKHQH[WWHQ\HDUVDQG what concretely should be prioritized in a phased, development plan.178 In 2003, the majority of roads were not paved, streetlights were in bad shape, the drainage system outdated, receipt collection, budget planning and procurement were weak, and trash collection was a major sanitation and health challenge.179 The first proposal for PROMAPUTO was presented to the stakeholders in May 2006. By

177 In conjunction with his town counselors, using the same sort of consultative model which he had used in vetting his 2003 Electoral Manifest, Dr. Comiche led three efforts²PROMAPUTO, the formation of the PEU, and the passing of the participatory budget process²which provided the CMM the foundation for its economic and community development not just during his mandate but for years to come. Approaching the World Bank, he proposed that some of the leftover funding from a previous World Bank project be used to undertake a massive consultation process in which the city would construct a broad-based, multi-year development plan. What came of it was not only the PROMAPUTO strategy but also the Urban Structure Plan or Plano de Estrutura Urbana (PEU), published in DQGDSSURYHGE\WKH0XQLFLSDO$VVHPEO\LQWRZDUGVWKHHQGRI'U&RPLFKH¶V mandate.

178 $OWKRXJK'U&RPLFKHDFNQRZOHGJHVWKDW³3520$3872GLGQRWVSURXWIURPP\KHDG´KHZDV key to the consultation process itself and procured the initial funding used to conduct such an extensive process. Author interview, Dr. Eneas Comiche, Maputo, December 12, 2009.

191

January, 2007, it was approved by the World Bank. Dr. Comiche ended a subsequent presentation

VWDWLQJWKDWDSULRULW\IRUWKH&00ZDVWRDFWLQWKH³VSLULWRILPSURYLQJVHUYLFHVWRWKH0DSXWR

UHVLGHQW´ZKLFKLQYROYHGLPSURYHPHQWVLQWUDVKFROOHFWLRQDQGPDUNHWLQIUDVWUXFWXUH

The planning exercise for PROMAPUTO led to an important legacy planning project--

0DSXWR¶VILUVWSDUWLFLSDWRU\EXGJHWSURFHVV180 The purpose of enacting a participatory budget process

ZDVWRFRPSO\ZLWKRQHRI'U&RPLFKH¶VHOHFWRUDOSURPLVHVRI³SULYLOHJLQJDWUDQVSDUHQWJRYHUQPHQW established in popular consultation and the involvement of all municipal residents in decision-making

SURFHVVHV´7KH)<%XGJHWZDVWKHILUVWZKLFKXWLOL]HGWKLVPHWKRGWRDOORFDWHLQIUDVWUXFWXUH transfers (FIIL) to Maputo.181 Independent newspaper Zambeze reported that Dr. Comiche, unlike his

SUHGHFHVVRUV³JDYHIOLJKW´WRXUEDQGHYHORSPHQWSURJUDPVZKLFKLQFOXGHGXSGDWLQJWKHGUDLQDJH system, constructing new roads, and improving trash collection (Mate 2008,14).

$QRWKHUH[DPSOHRIKRZ'U&RPLFKH¶VOHDGHUVKLSVW\OHVXSSRUWHGLPSURYHPHQWVLQPXQLFLSDO

VHUYLFHZDVWKH&00¶VSDUWQHUVKLSZLWK$*5(687KLVSDUWQHUVKLSQRWRQO\LPSURYHGWUDVK collection in the shorter term but also revenue generation for longer-term sustainability of services. The

179 One assessment of city infrastructure indicated that sanitation and drainage services only covered 30 percent of the urban zone and 20 percent of the suburban zone.

180 The purpose of the PEUMM was to update and articulate an urban land use and infrastructure plan that could accommodate growth in the periurban areas, establish city boundaries and plan for the maintenance and expansion of infrastructure. Under the guidance of Dr. Comiche, two committees constructed the PEUMM-- one comprised of town councilors and the other was a consultative committee comprised of the rest of the town councilors, private sectors participants and representatives of surrRXQGLQJGLVWULFWVWKHSURYLQFHDQGFLW\RI0DWROD$IWHUWKH&00¶VWHFKQLFDOUHYLHZDQG presentation to the two committees, public hearings of the PEUMM were held over three days in all seven districts at which 750 members of the public attended. Another series of public hearings were held before the plan was finalized in September, 2008 and presented to the Municipal Assembly for approval.

181 7KLVSURFHVVZDVEDVHGRQWKH0D\RU¶VRQJRLQJFRQVXOWDWLRQSURFHVV(DFKRIWKHWRZQ councilors nominated by the Mayor is in charge of at least one bairro and is charged with coordinating hearings with the bairro leadership, namely the Neighborhood Secretary (Secretario do Bairro) and the EDLUUR¶V Consultative Council (Conselho Consultivo). The Activities Plan and Five-Year Plan were based on these consultations. In 2008, the CMM coordinated consultations based on this model to gather priorities for the 63 bairros of what investments they would like to see in their neighborhoods, which culminated in the FY 2009 Budget. Dondo is the only other municipality in the study, which employs a participatory budget model similar to the Maputo model.

192

&00¶VILUVWDWWHPSWWRFKDUJHIHHVIRUWUDVKVHUYLFHZDVPHWZLWKFRQVLGHUDEOHUHVLVWDQFHGXHWRSRRU

SODQQLQJ7KH*7=¶VAGRESU, or Apoio à Gestão de Resíduos Sólidos Urbanos (Support to Solid

Waste Management) helped the CMM design a sliding-scale fee structure for trash collection which is a national model today. While AGRESU had been providing technical assistance since the Cananá

$GPLQLVWUDWLRQLWZDVXQGHU0D\RU&RPLFKH¶VOHDGHUVKLSWKURXJKKLVWRZQFRXQFLORUIRUVDQLtation in which the majority of technical analysis, feasibility studies and eventual adoption of a realistic fee scale that increased fee revenue took place. The CMM first began charging a flat 20 MTn/month fee in 2002 through the electricity bill. The CMM had done no outreach to inform the population of this charge and, when coupled with no visible improvement to SWM service, residents simply refused to pay the fee in protest despite the risk of having their electricity cut off.

In response, the CMM, with the assistance of AGRESU, undertook two responses. First, the CMM conducted a public information campaign to explain the use of the fees. Second, in 2003, the CMM began experimenting with a scaled fee for domestic and commercial producers to replace the flat fee. It enlisted the technical assistance of AGRESU to work with the Directorate of Sanitation to determine the

RYHUDOOVWDWHRIWKH&00¶VWUDVKFROOHFWLRQFDSDFLW\ SHUVRQQHODQGHTXLSPHQW ZKHUHWKH&00 needed to build that capacity and how to finance the gap with both donor funds and adjustments to the fee schedule. The idea was to design a viable market-based sliding scale fee schedule on which the

CMM could rely for both current and future revenue. In 2007, the CMM began charging a scaled fee after research done by AGRESU indicated that trash production is linked to energy consumption. The

CMM changed the fee rubric from a flat 20 MTn per family and 5 MTn for social tax to a fee scale based on energy consumption. In 2008, the Municipal Assembly codified the trash fee scale for its three categories of producers. Monthly collections more than doubled from 1,700,000 MTn under the old fee

VFKHGXOHWR07QXQGHUWKHQHZRQH ³3ODQR'LUHFWRU´, 33-34).

193

Beira

Simango inherited a city where the cholera fatality rates at one point were almost twice as high

DVWKHQDWLRQDODYHUDJH ³&KROHUD'UDPDLQ%HLUD´,10).182 The city was regularly flooded and trash collection was so poor that mounds of trash blocked streets. Simango became known as

Chinamiru CheduZKLFKLQ1GDXWUDQVODWHVWR³WKHRQHZKRXQLWHVXV´DKRPHWRZQKHUREHFDXVH since the beginning of the mandate, the conditions visibly improved. However, unlike Maputo, improvements seemed tied more to Simango than to a sustainable system of modern public administration. I attribute this primarily to the fact that Beira is an opposition municipality, which creates a volatile political environment, making consolidation in performance gains in the form of modern public administration difficult.

In 2003, Beira (CMB) did not have functioning street lamps and computers, where they existed, were completely obsolete. Like the other cases, Simango and his appointees came into offices with little to no furniture and municipal personnel were only sporadically paid. When the FRELIMO party found out they were out of power, municipal staff racked up even more municipal debt.183 Market fee collectors simply put fees in an envelope and left the envelope in a box in the municipality, a practice open to abuse. The CMB collected delinquent taxes by placing ads in the newspaper to request both individuals and corporations to pay back taxes which little success.

6LPDQJRLQKHULWHGZKDWKHFDOOHGD³VLVWHPDILQDQFLHURGHURER´(a financial system of robbery).184 Salary payments lagged three to four months behind and fees and taxes simply were not

EHLQJFROOHFWHGFRQVLVWHQWZLWKWKH&0%¶VDXWKRULWLHV185 With the assistance of trusted advisors, such as his cousin, and two long-time city employees, one of whom eventually supported his 2008 bid as an

182 During an epidemic in early 1998, the cholera fatality rate in Beira was as high as 5.6 percent, almost twice that of the national average of between 2 to 3 percent.

183 Author interview, Town Councilor for Finances, Beira, March 18, 2009.

184 Author interview, Mayor Daviz Simango, Beira, April 4, 2009.

185 One journalist reported in an interview that the Port of Beira and Beira Airport under the FRELIMO Administration were not required to pay taxes.

194

independent for Beira Mayor, Simango went about making changes to the way City Hall was run which increased revenues and improved services. Fee collections became systematized and large contributors pursued. Instead of anonymously depositing market fees in a box at City Hall, Simango put his cousin in charge of the Directorate of Markets and Fairs to systematize and record the collection of market

IHHV³7KH&0%LVWKHILQDQFLDORZQHURQWKHJURXQG:HQHHGWRDFWLYHO\SXUVXHIHHV´KHVWDWHG186

Instead of anonymous deposits in City Hall, a municipal manager sets and monitors the daily fee collection goals of twenty to thirty fee collectors who must turn in collections twice a day to him which he then records in a ledger and deposits in City Hall. A new financing system called Proreceitas was created specifically for Beira to issue bills automatically, which also improved fee collection.187 One municipal staffer who had been working for the CMB for four years indicated that receipts have grown because of the capacity to control the influx of receipts and the volume of receipts charged, with own- source receipts growing from between 27 to 32 percent of the total budget to 47 percent of the total budget in 2009.188 This is partly due to the pursuit of tax payments from large businesses like the Beira

Port, Beira Airport, and the Mozambican Railway.

Simango also reformed the municipal civil service. When he came into office in February

2004, there had been a municipal salary strike as January salaries had not been paid. Simango began

February 6th and by February 8th, January salaries were paid.189 The town councilor for Human

5HVRXUFHVLQGLFDWHGWKDWVLQFHWKH&0%¶VFDSDFLW\WRSD\VWDIILQFUHDVHG190 Actual staff also increased. In 2004, the CMB had 900 functionaries and by 2009, it almost doubled to 1,500.191 By

2007, Professional Management Review (PMR), a South-African ranking firm that conducts telephone

186 Author interview, Former Town Councilor for Markets and Fairs, Beira, March 9, 2009.

187 Author interview, Chief of Receipts, Directorate of Finance, Beira City Hall, Beira, March 19, 2009.

188 Ibid.

189 Author interview, Mayor Daviz Simango, Beira, April 4, 2009.

190 Author interview, Town Councilor for Human Resources, Beira, April 1, 2009.

191 Ibid.

195

surveys to rank businesses and government services annually, ranked Simango as one of the two best performing mayors in Mozambique (PMR Online 2008).192

7UDVKFROOHFWLRQRQHRI6LPDQJR¶VFDPSDLJQFRUQHUVWRQHVLPSURYHGDOPRVWLPPHGLDWHO\ once he was in office. He came into office in February 2004 and within a few months, piles of trash, which KDGOLQJHUHGRQWKHURDGIRU\HDUVGLVDSSHDUHG2QHIRFXVJURXSSDUWLFLSDQWLQGLFDWHGWKDW³ZH

XVHGWRGLHOLNHIOLHV´193 When asked how he was able to reverse trash collection so quickly, Simango stated he motivated the workers, focused on creating minimum conditions, which included paying them on time, and pressed the CMB to collect taxes and fees, which he then used to finance trash collection.

In 1998, 700 new cases were diagnosed during the peak of cholera season with a mortality rate twice the

QDWLRQDODYHUDJHLQ%HLUD SHUFHQWYHUVXVWRSHUFHQW ³&KROHUD'UDPDLQ%HLUD´,10).194

By the 2010 cholera season, the number shrank to 200, or a fatality rate of 0.3 percent, which was just undeUIRXUWLPHVWKHQDWLRQDODYHUDJH SHUFHQW ³,QIRUPHWULPHVWUDO´,2). One resident and self-proclaimed FRELIMO supporter summed the changes in Beira under Simango best:

I have been living in Beira for many years now. The first local governing authority in charge of running the city of Beira was the party in power, FRELIMO. They failed tremendously in their PLVVLRQWREHWWHUWKHORFDOV¶OLYHV7KHLUIDLOXUH V LQFOXGHG QRW SURYLGLQJFLW\FOHDQLQJ ordering of the urban spaces, control of the markets, no transparency in public funds management. Eventually the opposition took over. With their energized president Daviz Simango they are caring for all the basics and even doing more. Although I'm a FRELIMO supporter, I am forced to recognize their achievements (Posse 2006).

However, Beira, as the main opposition city is constantly under scrutiny by the media or under

LQYHVWLJDWLRQ)LUVW6LPDQJR¶VWRSVWDIIDSSHDUWREHFKRVHQPRUHIRUOR\DOW\WKDQWHFKQLFDOH[SHUWLVH because, in the politically charged environment of Beira politics, he arguably needs to be able to trust his managers.195 )RUH[DPSOHWKHIRUPHUWRZQFRXQFLORUIRU0DUNHWVDQG)DLUVLV0U6LPDQJR¶V

192 Dr. Eneas Comiche of Maputo was the other mayor.

193 Participant, Palmeiras I focus group, Beira, March 17, 2009.

194The peak was January-February, which are the hot humid months of the year. The season typically is between January and March.

195Although Dr. Comiche, like all mayors, can nominate up to 50 percent of his town councilors from any sector using his own criteria, the town councilors chosen tended to be experienced in their fields. The Sanitation and Salubrity town councilor had worked with the National Aids Council, started his 196

FRXVLQ$OWKRXJK6LPDQJR¶VFRXVLQVXFFHVVIXOO\OHGDPDMRUUHQRYDWLRQRIRQHRIWKHODUJest markets in

Beira, he was forced to resign from the town councilor position due to accusations that he had stolen market fees. 196 6HFRQGDUHYLHZRISXEOLFPXQLFLSDOSHUVRQQHOUHFRUGVUHYHDOHGPRVWRIWKH&0%¶V personnel, especially his top advisors, do not go through the central government appointment process.197

Unlike the other cases, the majority of municipal personnel I interviewed were appointed during the

Simango administration.

Lastly, as the 2008 elections drew near, rumors circulated about potential FRELIMO sabotage.

Market fees were calculated by hand because the Markets and Fairs computers had been stolen, resulting in the loss of all digitized revenue records. One senior official speculated that the computers where stolen by FRELIMO operatLYHVWRVHHZKDWWKH&0%¶VUHYHQXHKDGEHHQLQPDUNHWIHHVWR

SHUKDSVGLVFUHGLW6LPDQJR¶VFODLPVWRLQFUHDVLQJIHHUHYHQXHVGXULQJWKHPD\RUDOHOHFWLRQV

6HFRQGRQHRI6LPDQJR¶VWUXVWHGDGYLVRUVKDGDFRPSXWHUVWROHQVRPHVSHFXODWHDJDLQE\)5(/,02 operatives to gain insider campaign information. This advisor was then deported because he was a

Zimbabwean citizen who allegedly was in the country illegally. Perhaps a town councilor summarized

WKHIDFWRUVWKDWH[SODLQ%HLUD¶VG\QDPLFEHVWZKHQKHVDLG³:HDUHLQWKHGHYHORSLQJZRUOGSRRUDQG

LQWKHRSSRVLWLRQ:HKDYHWRGRVRPHWKLQJIRURXUVHOYHV´198

own NGO and worked for the municipality for many years but did not have specific technical experience in the area of sanitation or health. In fact, his NGO, the Group of Reflection and Change (Grupo de Reflexão e Mundança) nominated Mr. Simango for his 2008 re-election campaign after Mr. Simango was ejected form RENAMO, as required by law.

196 Author interview, Former Town Councilor for Markets and Fairs, Beira, March 9, 2009.

197 City employees fall into two categories: inside the framework (dentro do quadro) and outside the framework (fora do quadro). Employees inside the framework are nominated by the municipality and DSSURYHGE\WKH0LQLVWU\RI6WDWHDQG$GPLQLVWUDWLRQDQGWKH,QVSHFWRU*HQHUDO¶V2IILFH7KHSULPDU\ benefits associated with this status are that employees qualify for national retirement benefits and promotions are easier to get. Employees outside the framework are municipal staff who are hired in theory as temporary workers on short-term contracts. These workers do not have a right to national retirement benefits are have less job security. Review of the municipal 2009 Plan of Activities and Budget indicated that the majority of 1,600 municipal employees did not go through this appointment process (indeed most did not have a tax identification number or NUIT).

198 Author interview, Town Councilor of Sanitation, Beira, March 12, 2009.

197

Dondo

Manuel Cambezo is one of five mayors who has been consecutively re-elected since 1998.199

His background in municipal government and the consistency of his management team have combined to provide a high level of service in the second mandate. Under Cambezo, Dondo has been held up as a

QDWLRQDODQGLQWHUQDWLRQDOVWDQGDUGLQLPSOHPHQWLQJSDUWLFLSDWRU\JRYHUQPHQW ³.H\&RPSHWHQFLHV´

2005) ³0XQLFLSDO'HYHORSPHQWLQ0R]DPELTXH´,6). Furthermore, Cambezo is responsible for a number of major infrastructure investments in Dondo. The CMD constructed a new market, the largest in the municipality, for vendors, which they subsequently named after the mayor in gratitude. He was also responsible for the new conference center, which houses both his office and Municipal Assembly proceedings. Since 1999, water has been available 24 hours a day and electricity has become selectively available.200 Tractor service for trash pick-up became regular, if not frequent, in informal areas, and the

CMD installed water fountains in these same areas. Like Beira and Maputo, Dondo has set up a low cost mapping and property registry on which it can build its nascent SURSHUW\WD[V\VWHP ³0XQLFLSDO

'HYHORSPHQWLQ0R]DPELTXH´, 'XULQJ&DPEH]R¶VWHQXUH'RQGRVHUYHGDVWKHKRVWIRU

0R]DPELTXH¶VILUVWUHXQLRQRIDOOPD\RUVWKURXJKWKH1DWLRQDO$VVRFLDWLRQRI0R]DPELFDQ

Municipalities (ANAMM).

However, despitHWKHDFFODLPRI'RQGR¶VSDUWLFLSDWRU\EXGJHWSURFHVVWKHLPSOHPHQWDWLRQRI community hearings is weak. One study found that the process was somewhat shallow. In discussions with the development councils in each neighborhood, issues of budgeting typically do not surface, confining discussions to project wish lists (water fountains, roads, etc) (Nuvunga, Mosse, Varela

2007,11-14). Vertical accountability mechanisms, which residents can use to compel the CMD to follow-up on projects were found to be weak. Another study commissioned by the Austrian

Development Cooperation, the donor which funded this process initially in Dondo, raised concerns about the sustainability of implementation (Manndorff, Nöst et al 2004,4). Although financial

199 The other four municipalities are Montepuez, Quelimane, Manica, and Maxixe.

200 Participant, Formal neighborhood focus group, Dondo, November 9, 2008.

198

management of donor projects was transparent, other municipal expenditures were not consistent with existing plans

I attribute the lack of depth of participatory budgeting but progress in municipal infrastructure and service to a management team with years of experience, which has not been challenged politically to compel them to meaningfully share power. Cambezo and his management team have several years of experience. Cambezo was the head of the Beira government in the 1980s, prior to decentralization.

He had the technical background and the backing of the FRELIMO party as the former administrator in

Beira and two-term mayor in Dondo. As a result, Cambezo dominates the CMD, marginalizing the

Municipal Assembly. For example, CMD proposals and budgets are supposed to be sent over 12 to 15 days before the start of the legislative session; however in Dondo, the Municipal Assembly received proposals only three hours before discussions (Nuvunga, Mosse, Varela 2007, &DPEH]R¶V management team members have all been with the CMD since 1998. The town councilors for solid waste management and local economic development have been in those positions since 1998. While the

Finance town councilor came into this position in 2003, he had been with the CMD since 1998. While local researchers have observed a lack of vertical accountability between the mayor and the Municipal

$VVHPEO\DFRPPRQIHDWXUHRIPXQLFLSDOJRYHUQPHQWLQ0R]DPELTXH&DPEH]R¶VH[SHULHQFHDQGWKH

FRQVLVWHQWOHDGHUVKLSRIKLVPDQDJHPHQWWHDPUHIOHFWLQWKH&0'¶VSHUformance.

2008 Mayoral Elections

Maputo

Despite successfully pursuing anti-corruption measures introduced in his 2003 FRELIMO

Electoral Manifest, Dr. Comiche was not re-nominated by FRELIMO for a second mandate because fulfilling campaign promises of anticorruption and better municipal performance interfered FRELIMO party interests. He was often seen as a party outsider because, in an effort to enact municipal regulations and bring professionalism to the work of the CMM, he limited party largesse. Dr. Comiche 199

ZDVNQRZQDV³LQFRUUXSWLEOH´PDNLQg the budget visible and addressing corruption head on by ending

WKH³mor do´V\VWHPRUWKH³ELWH´+RZHYHUWKLVZDVHQRXJKWRGLVJUXQWOHDODUJHJURXSRIEHQHILFLDULHV of the mor do V\VWHPZKHUHWKHFLW\ZDVVHHQDVWKHLU³vaca leteira´RUPLONLQJFRZ

Comiche administration missteps with FRELIMO-connected businesses did not go unnoticed.

7KH)5(/,02'HSXW\LQWKH0XQLFLSDO$VVHPEO\LQGLFDWHGWKDW³LQIDFWWKHUHZHUHVRPHHUURUVLQ

GHDOLQJZLWK WKHEDVH´VSHDNLQJRI&RPLFKH¶V$GPLQLVWUDWLRQ 0RTXLYDODND 2008). Examples cited in the press include the clashes with profitable business ventures connected to FRELIMO party members such as demolishing a profitable club in a high-income area due to its lack of abiding by municipal regulations. Another incident UHSRUWHGLQWKHSUHVVSHUWDLQHGWR0D\RU&RPLFKH¶VGHQLDORI0XQLFLSDO

Assembly member requests for access to land originally allocated to families displaced by the explosion

RIPXQLWLRQVLQDZDUHKRXVHLQ0DOKD]LQH ³(QHDV&RPLFKH´ 0D\RU&RPLFKHUeviewed the request and indicated that there was no provision in the law to allow for such a land transfer. Although it would have benefitted the entire Municipal Assembly, which includes three political parties, 78 percent of the seats are FRELIMO-held, so this denial was registered wryly among some party members

DVDQH[DPSOHRIWKH³LQGHSHQGHQFHDQGYLVLELOLW\´ZLWKZKLFK0D\RU&RPLFKHDFWV ,ELG

By the summer of 2008, grumblings of a forced retirement for Dr. Comiche in favor of the

Minister of Sports and Youth and former Governor of Niassa province under President Chissano, David

6LPDQJREHJDQWRFLUFXODWH ³)5(/,026WULYLQJIRU6WURQJ0XQLFLSDO&DQGLGDWHV´ 201 Dr.

&RPLFKHLQGLFDWHGWKDW'DYLG6LPDQJR¶VFDQGLGDF\³GRHVQ¶WWKUHDWHQPLQH´DQG planned to continue to

YLHIRUWKHSDUW\¶VQRPLQDWLRQIRUDVHFRQGWHUP (Nhamirre and Sambo 2008). However, according to

201 Despite talk of nominating candidates based on performance, the press reported that FRELIMO ZRXOGZDVHQWHUWDLQLQJRWKHULQGLYLGXDOVLQ0DSXWR2QHQHZVVRXUFHVWDWHGWKH(QHDV&RPLFKHLV³E\ no means guaranteed of becoming the FRELIMO candidate, despite his universally acknowledged success in improving lLYLQJFRQGLWLRQVLQWKHFLW\RYHUWKHSDVWILYH\HDUV´'U&RPLFKHJDYHDQ interview in which he acknowledged a divide within the FRELIMO party between an old and a new JHQHUDWLRQLQZKDWZDVGXEEHGDVDQ³RSHQZDU´7KHQHZJHQHUDWLRQVXSSRUWHG'U&RPLFhe and SRLQWHGWRWKHFHPHQWFLW\¶VLPSURYHGLQIUDVWUXFWXUHDQGVHUYLFHVDVSURRIRIZK\KLVFRQWLQXHG candidacy is merited whereas the old guard were dissatisfied with the progress the Comiche Administration had made in the peripheral shantytowns and advanced David Simango as the alternative.

David Simango, current Mayor of Maputo, and Daviz Simango, current Mayor of Beira, are not related.

200

one prominent opposition politician, Dr. Comiche was part of the Chissano wing whereas Simango was a part of the Guebuza wing, which some felt was ascendant. On August 22, 2008, after 13 hours of deliberation, the President of the FRELIMO Committee announced that the winner of the internal election for the FRELIMO ticket in the third municipal elections was David Simango, not Dr. Eneas

Comiche. Of the three candidates put forth in the internal vote, Dr. Comiche came in second with only

32 percent of the vote (or 25 votes) with David Simango winning over two-thirds of the Committee vote

(or 53 votes) (Manguana 2008). Dr. Comiche accepted the outcome, despite his performance and popularity among voters, and did not seek to create another party or run as independent, likely due to his

OR\DOW\WRDQGVWUHQJWKRIWKHGRPLQDQWSDUW\ ³2SLQLRQ3ROOV´ 202 David Simango won the 2008 mayoral election, winning 85.78 percent of the vote (see Table 6.5).

Beira

%HLUD¶VPD\RUDOQRPLQDWLRQVDQGHOHFWLRQVSURFHVVSURYLGHGHYLGHQFHRIKRZPD\RUDO performance encouraged political competition at the local level, a dynamic absent from Maputo elections. First, weakness and disorganization in the opposition party created space for a strong candidate to break through and base his successful candidacy on performance. Second, that candidate²

Daviz Simango²was able to capitalize on both the anti-FRELIMO and anti-RENAMO sentiments to create a new party, which, while limited, won representation in the Parliament in the 2009 national elections.

By the summer of 2008, RENAMO party leader Afonso Dhlakama had stated that RENAMO would back all currently-standing mayors. In a sudden departure from stated policy, RENAMO leadership announced that it would not be re-nominating Daviz Simango for a second term.203 This

202 A pre-election opinion poll of 2,000 Maputo residents indicated that 31.9 percent would vote for Dr. Comiche, more than twice the number of the eventual FRELIMO nominee, David Simango (12.05 percent). Almost 60 percent (59.85 percent) of those surveyed said that they would vote for Dr. &RPLFKH ³GRQ¶WNQRZ´DQG³ZRQ¶WVD\´UHVSRQVHVZHUHUHPRYHG

203 ³2SUHVLGHQWHGRPXQLFtSLRGD%HLUD'DYL]6LPDQJRIRLKRMHH[SXOVRGD5(1$02SULQFLSDO partido da oposição moçambicana, por se candidatar como independente às eleições autárquicas de 201

announcement was made four days before the deadline for registering candidates after the Simango campaign spent funds on advertisements. Mayor Simango decided then that he would not run for

Mayor; however, there was pressure on him to run from the RENAMO party bases and he decided to run as an independent candidate.204 As a result, Simango announced his intent to run. RENAMO

OHDGHUVKLSEDFNHG0DQXHO3HUHLUDIRU%HLUD0D\RUDQGH[SHOOHG6LPDQJRIURPWKHSDUW\ ³0D\RURI

%HLUD([SHOOHGIURP5(1$02´ )XUWKHUPRUH5(1$02WULHGWRQXOOLI\KLVLQGHSHQGHQW candidacy by appealing to the Ministry of State Administration (MAE); however, this request was unsuccessful.

Although the FRELIMO candidate, Laurenço Bulha, had far more publicity than the other five

FDQGLGDWHVLWZDVHYLGHQWWKDW'DYL]6LPDQJRZDVWKHIDYRULWH ³)5(/,02&KRRVHV,WV6RIDOD

CandidaWHV´ 205 He ran on a results-EDVHGSODWIRUP³(OH3URPHWHH)D]´RU³+HSURPLVHVDQG

GHOLYHUV´ZDVKLVFDPSDLJQVORJDQ7KHODVWZHHNHQGEHIRUHWKHHOHFWLRQFDUDYDQVRIWKHWRSWKUHH candidates²Simango, Bulha and Pereira²circulated through the city and this observer noted that the caravan for Mayor Simango was the longest, with Pereira barely making a showing. RENAMO was so unprepared for this election that Pereira did not even have a 2008 Electoral Manifest. Simango won by a landslide²over 61 percent²almost twice that of the FRELIMO candidate, Laurenço Bulha (33 percent), and over twenty times that of the RENAMO candidate, Manuel Pereira (just under 3 percent)

(see Table 6.4). When Simango won, he purchased beer and food for block parties all over the city and published a thank you note to his constituents for having voted for him. RENAMO suffered a humiliating national defeat, having not only lost all five municipalities it gained in the 2003 municipal elections but also not gaining any of the ten new municipalities. This left FRELIMO in control of 42 of

WKHPXQLFLSDOLWLHVDQGSURYLGLQJ5(1$02¶V IRUPHU VWURQJHVWFDQGLGDWH²Daviz Simango in

1RYHPEURDQXQFLRXR&RQVHOKR1DFLRQDOGRSDUWLGR´6HSWHPEHU$FFHVVHGRQ$XJXVW 2011 at http://ww1.rtp.pt/noticias/?article=149633&visual=3&layout=10

204 With three days to go, his campaign focused on gathering the requisite 3,000 signatures needed to be able to run as an independent. In that period of time, Simango received 11,000 signatures.

205 %XOKDZDV)5(/,02¶V6RIDODSURYLQFH)LUVW6HFUHWDU\DQGUDQIRUWKH)5(/,02QRPLQDWLRQIRU mayor of Beira unopposed.

202

Beira²with the platform from which to construct a party that would further humiliate Dhlakama in the

2009 national elections.

$IWHU6LPDQJR¶VYLFWRU\KHVHQWDGYLVRUVWR5(1$02OHDGHUVKLSWRUHFRQFLOH &KLFKDYD

2010,8-9).206 According to Simango sources, Dhlakama rejected these advances. Meanwhile, neighborhood-level brigades in various provinces respondeGWR6LPDQJR¶VYLFWRU\DJDLQVWQRWRQO\

FRELIMO hegemony but also challenging the notion that RENAMO was the only alternative. The

Simango campaign apparatus eventually became the staff and headquarters for the MDM, or the

Democratic Movement of Mozambique (Movimento Democrático de Moçambique). This new

PRYHPHQWDUJXHGWKDWEHFDXVHRI5(1$02¶VDE\VPDOVKRZLQJLQWKHPXQLFLSDOHOHFWLRQVWKHUH was no longer a viable opposition party in Mozambique to challenge FRELIMO in the upcoming 2009 presidential DQGQDWLRQDOSDUOLDPHQWDU\HOHFWLRQV7RVLJQDOWKDWWKLVODFNRIRSSRVLWLRQZDVD³ZDNH-up

FDOO´WRRUJDQL]HWRVDYH0R]DPELFDQPXOWLSDUW\GHPRFUDF\WKH0'0FKRVHDURRVWHUZLWKDVXQULVLQJ in the background as its party symbol.

The MDM had its first congress to elect party leadership in March 2009 at the historic Catholic

University of Beira, a university, which was a concession in the Rome Peace talks between RENAMO and FRELIMO. There was an average of 20 delegates per province, with the room containing approximately 300 attendees. Representatives from prominent opposition parties from all over the

ZRUOGVXFKDV7VYDQJDUL¶V0RYHPHQWIRU'HPRFUDWLF&KDQJHDQG3RUWXJDO¶V6RFLDO'HPRFUDWLF3DUW\ were in attendance. Interspersed with the national anthem and Ndau songs, were speakers who made

UHIHUHQFHWRWKHQ'HPRFUDWLF3UHVLGHQWLDOFDQGLGDWH%DUDFN2EDPD¶V³\HVZHFDQ´VWDWLQJWKDW³%HLUD

LVWKHHSLFHQWHURIGHPRFUDF\´DQG³WKHRQHDQGRQO\RSSRUWXQLW\IRUWKHFRXQWU\WRUHWXUQWRD multiparty V\VWHP´6LPDQJRFORVHGWKH0'0¶VILUVW&RQJUHVVE\VWDWLQJWKDWWKH0'0ZDVSUHSDUHG

206 Given not only his victory but the slim margin the RENAMO party had in the Beira Municipal $VVHPEO\ SHUFHQWWR)5(/,02¶VSHUFHQW DQGFDWDVWURSKLFORVVHVDOORYHUWKHFRXQWU\6LPDQJR thought Dhlakama would surely make peace. Furthermore, Sofala province, of which Beira is the capital, was, up until that point, a RENAMO stronghold, with RENAMO winning 76.8 percent of the vote in 1994, 71 percent in 1999 and 65 percent in 2004 in national elections. By 2009, FRELIMO gained inroads in Sofala, winning, for the first time since the first national elections, the majority of votes (164,100), with the MDM coming in second (75,732) and RENAMO third (74,734).

203

WR³ILJKWIRUWKLV\RXQJGHPRFUDF\´LQDZD\WKHZHDNHQHG5(1$02FRXOGQRORQJHUFUHGLEO\GR

(Simango 2008,8).

:LWK6LPDQJRDVWKH0'0¶VFDQGLGDWHIRUSUHVLGHQWWKH0'0headed into the 2009 elections ready for a dog fight. Controversy followed when the National Elections Commission disqualified

MDM candidates from running in seven of the eleven provinces based on irregularities, allowing MDM candidates only in Maputo, Inhambane, Sofala and Niassa provinces (Hanlon 2010,8). 207 Although the

MDM was unable to compete in all provinces, it still won 24 seats in the Parliament.208 Despite

*XHEX]D¶VODQGVOLGHYLFWRU\6LPDQJRZRQSHUFHQWRIWKHYRWHWKHWKLUGODUJHVWSDUW\ZLn,

³HVWDEOLVKLQJKLPVHOIDVWKHILUVWFUHGLEOHRSSRVLWLRQFDQGLGDWHIURPD\RXQJHUJHQHUDWLRQ´DFFRUGLQJWR long time Mozambique political analyst, Joseph Hanlon (Hanlon 2009,1-2).209 'KODNDPD¶VVKRZLQJDW just 16 percent of the vote in 2009, a little more than half of the votes he received at his peak in 1994

SHUFHQW UHSUHVHQWHGWKHORZHVWSRLQWLQ5(1$02¶VHOHFWRUDOKLVWRU\ ,ELG, 2). Despite the fracturing effects the MDM had on dividing opposition votes against FRELIMO and the weaknesses in leadership and articulation of a viable alternative platform to FRELIMO, the MDM won inroads in

FRELIMO-dominated Maputo by gaining three seats in the Parliament²something that had never occurred in the history of Mozambican multiparty politics (Chichava 2010,10).210

207 Although the MDM challenged this decision and lost, the National Elections Commission (CNE) inadvertently included a candidate from another party (PDD) on a provincial list, which the PDD did QRWVXEPLWIRUFRQVLGHUDWLRQLQWKLVGLVWULFW$VDUHVXOWWKH3''FDQGLGDWH¶VQDPHZDVRQWKHEDOORW and he won the seat. The CNE invalidated the election but was forced to admit that its internal control WDEOHWKHVDPHRQHZKLFKWUDFNHG³LUUHJXODULWLHV´ZLWK0'0FDQGLGDWHVZDVIDXOW\

208 5(1$02REWDLQHGRQO\VHDWVZLWK)5(/,02ZLQQLQJOLRQ¶VVKDUHVHDWV&KLFKDYD

209 Guebuza won 75 percent of the vote and Dhlakama won only 16 percent in his fourth run as the RENAMO presidential candidate.

2106pUJLR&KLFKDYD³0'0DQHZSROLWLFDOIRUFHLQ0R]DPELTXH"´,QVWLWXWRGH(VWXGRVVRFLDLVH (FRQyPLFRV ,(6( Maputo, Mozambique for the conferencHµ(OHFWLRQSURFHVVHVOLEHUDWLRQ PRYHPHQWVDQGGHPRFUDWLFFKDQJHLQ$IULFD¶0DSXWRǦ 11 April 20101: 10, Accessed on August 15, 2011 at http://www.iese.ac.mz/lib/publication/proelit/Sergio_Chichava.pdf

204

Dondo

The 2008 Dondo mayoral contest provided another example of the strength of FRELIMO

SHUIRUPDQFHDQG5(1$02SROLWLFDOGLVRUJDQL]DWLRQ ³)5(/,02&KRRVHVLWV6RIDOD&DQGLGDWHV´

2008). From 2003 to 2008, Cambezo built a new market, conference center, regularized trash service and improved the reliability and availability of water and electricity. Although informal neighborhoods and markets still suffered from a comparable lack of municipal service relative to formal neighborhoods

DQG'RQGR¶VPuch touted participatory budgeting process appeared to be shallow by some local observers, Cambezo was still credited for improving service in Dondo. Dondo is a FRELIMO- stronghold in the predominantly RENAMO-dominated province of Sofala. Coming from what amounted to a surprising defeat in 2003 with the loss of five municipalities, one of which is located one hour from Dondo (Beira), FRELIMO had its eye on the prize of installing FRELIMO mayors and

PXQLFLSDODVVHPEO\PDMRULWLHVLQDOOPXQLFLSDOLWLHV ³Partido FRELIMO reitera expectative de

YLWyULDQDVDXWDUTXLDVGH6RIDOD¶´,4). This meant that Dondo had special symbolic significance because FRELIMO could not afford to lose the municipality located so close to Beira. As a high- performing mayor with WZRFRQVHFXWLYHWHUPV&DPEH]RSURYLGHGDFRXQWHUEDODQFHWR%HLUD¶V6LPDQJR

Cambezo was re-QRPLQDWHGZLWKQRRSSRVLWLRQ'RQGR¶VSHUIRUPDQFHZDVWKHFRUQHUVWRQHQRWMXVWRI

&DPEH]R¶VFDPSDLJQEXWDOVRIRUWKHRWKHUIRXUPXQLFLSDOLWLHVLQ6RIDODSURYLQFH. One FRELIMO

SDUW\UHSUHVHQWDWLYHVWDWHGWKDWZLQVLQIRXU6RIDODPXQLFLSDOLWLHVDUHEDVHGRQ³JRRGZRUN´OLNHWKDW done in Dondo which is the reason why residents will want Cambezo, for example, to continue (Ibid,4).

The political controversy in Beira sSUHDGWR'RQGR'DYL]6LPDQJR¶VIRUPHU)LQDQFHWRZQ councilor, Manuel Bissopo, was one of the three town councilors Simango fired in September because

WKH\VXSSRUWHG'KODNDPD¶VGHFLVLRQWRZLWKKROGEDFNLQJIURP6LPDQJR 'DYL]6LPDQJR([RQHUD7UrV

Vereaçoes" 2008).211 One well-respected independent newspaper reported that instead of focusing on his two rivals or Dondo issues, Bissopo instead disparaged Simango in Beira. Additionally, the validity

RI%LVVRSR¶VRZQFDQGLGDF\ZDVFKDOOHQJHGPDNLQJ5(1$02ORok even less credible. The National

211 The other two town councilors were Fernando Mbararano (Construction and Urbanization), a RENAMO provincial delegate, and Faque Ferraria Inácio (Civil Protection and Transportation).

205

Elections Commission (CNE) disqualified Bissopo because he did not meet Dondo residency requirements to run as mayor, a decision in which Manuel Cambezo was implicated for interfering in the provision of residency documeQWV 2V+iELWRV'H0XJDEH´,4).212 Regardless, RENAMO

FRQWLQXHGWREDFN%LVVRSRDVWKHLUFDQGLGDWH ³0DQXDO%LVVRSRQDµFDoD¶DRYRWR´, 2). The CNE decision was eventually overturned by the Constitutional Council four days before the close of the

RIILFLDOFDPSDLJQVHDVRQ ³7UrVFDQGLGDWRVUHTXDOLILFDGRVQR'RQGR*RURJRVDH0DQLFD´,2).

Then, once Bissopo was able to re-ODXQFKKLVFDPSDLJQKHZHQWULJKWEDFNWR³ILJKWLQJ'DYLG

6LPDQJR«LQVWHDGRIFOHDUO\ZRUNLQJRQWKHEDVHVLQ'RQGR´ 1DWDO08,5).

Manuel Cambezo ran a confident 2008 campaign, having been re-nominated unopposed.

)5(/,02³KHDY\KHDYLHV´ZKRFDPSDLJQHGIRUKLPLQFOXGHGWKH0LQLVWHULQWKH3UHVLGHQF\RI

3DUOLDPHQWDU\$IIDLUVDQGPHPEHURI)5(/,02¶V&HQWUDO&RPPLWWHHDQGWKH Minister of Health

³'RQGRGRLVFDQGLGDWRUHVUH]DPDQWHVGRHQFDUUDPHQWRGDµFDoD¶´,4). One national newspaper

FRQGXFWLQJLQWHUYLHZVRI'RQGRUHVLGHQWVUHSRUWHGWKDW'RQGRKDV³LPSURYHGZLWKHDFKSDVVLQJ\HDU´ and point to the improvement in water distribution and paved streets (Mavila 2008,25). Cambezo won

KLVWKLUGWHUPZLWKWKHGRPLQDQWPDMRULW\RIYRWHV SHUFHQW WR%LVVRSR¶VSHUFHQW VHH7DEOH

6.5).

Conclusion

Did elections work as a sanctioning mechanism to foster mayoral accountability and improve municipal performance? My findings indicate elections worked as a pressure on mayors to improve municipal performance to the extent that the voters were able to reward mayors in the 2008 election based on their compliance with 2003 campaign promises. In Beira and Dondo, elections worked as a reward for high performing mayors whereas in Maputo, voters did not have the option. I attribute variations in the efficacy of votes as a sanction (or reward) in demanding higher municipal performance

212 Bissopo was reported to be a resident of Dondo since January, thereby surpassing the six month residency requirement. Cambezo was accused of interfering in the process to grant him residency documentation to prove his ability to run in Dondo.

206

to the following two main factors: first, the ability to mobilize a management team that could implement visible performance improvements to deliver on 2003 campaign promises; and second, the strength of the nominating political party.

The third unanticipated finding, which was constant across all three municipalities, was that voter turnout for local elections as national elections turnout declined, suggesting that political decentralization does indeed promote democracy at the local level. This finding suggests that, in the long term, greater voter turnout in local elections might re-invigorate Mozambican politics by creating conditions for real political competition.

Technical expertise and mobilizing a competent management team was a constant factor across

DOOWKUHHPXQLFLSDOLWLHVWKDWKHOSHGWKHVHPD\RUVLPSOHPHQWWKHLUHOHFWRUDOPDQLIHVWV&RPLFKH¶V management team was comprised of outside experts who had no previous experience with municipal government. He chose doctors, businessmen, and finance experts based not only on their good standing

ZLWKKLPEXWWKHLULQGHSHQGHQWH[SHUWLVH6LPDQJR¶VWHDPRIWUXVWHGDGYLVRUVwas relatives, former

CMB employees, and RENAMO operatives who got the job done. Simango himself is a trained civil engineer. Cambe]R¶VWHDPZDVDOVRFRPSULVHGRISHUVRQDODFTXDLQWDQFHVZKRDIWHUDFRXSOHRI

PDQGDWHVKDGJDLQHGWKHH[SHUWLVHWRFRQWLQXHZRUNLQJXQGHU&DPEH]R¶VFRQVLVWHQWOHDGHUVKLS,QDOO three cases, the mayors were personally considered technically proficient and able to implement a consistent change in their respective municipalities.

7KHVWUHQJWKRIQRPLQDWLQJSROLWLFDOSDUWLHVSOD\HGDGHFLVLYHUROHLQHOHFWLRQV%HLUD¶V

2008 election process illustrated the weakenss of RENAMO. A lack of cohesive party leadership

XQGHUPLQHGWKHSDUW\ZLWKVRPHDUJXLQJWKDWWKHZLWKGUDZDORISDUW\VXSSRUWRI6LPDQJR¶V candidacy was motivated by a perceived political threat Simango posed to RENAMO leadership. This created an opening which made a difference in BeirD6LPDQJR¶VJUDVV-roots connections with bairro leadership formed the springboard for his successful independent candidacy and the creation of the

0'0DFRQGLWLRQODFNLQJLQ&RPLFKH¶VSROLWLFDOODQGVFDSH213 Although Comiche was a high-ranking

213 Bairro leadership played a critical role in the creation of the MDM. In one MDM meeting in the Munhava headquarters in Beira, GeraUGR&DUYDOKRRQHRI6LPDQJR¶VWRSSROLWLFDOVWUDWHJLVWV 207

member of the FRELIMO Central Committee and he had improved municipal services, he did not have

WKHJUDVVURRWVSROLWLFDOFRQQHFWLRQVWRERWKOREE\DJDLQVWWKH)5(/,02OHDGHUVKLS¶VGLVDSSRLQWPHQW with his anti-corruption stance and then subsequently support a potential independent nomination when

Comiche was cut loose. Furthermore, RENAMO did not have the same access to state largesse and resources as did the governing dominant party so the opposition party did not exert the same level of party discipline that FRELIMO arguably could. One editorial put it best:

In Maputo, E(neas) Comiche was clearly put to the side in what appears to be counter the ZLVKHVRIPDQ\+HUHDFWHGREHGLHQWO\« DQG «0DSXWRUHVLGHQWVOLPLWHGWKHLUGLVFXVVLRQRI the epicenter of the problem. In the city of Beira, it was the opposite: Beira residents, independent of party affiliation, demonstrated to the country that those who do good work should not be substituted! They loved their mayor and did not let the political factor surpass the technical factor (Guevane 2008,12)(Azagaia 2008,14).

6HFRQGZKHUHDV5(1$02¶VDSSHDODVWKH³SDUW\RISHDFH´PD\KDYHLQWKHSDVWDOORZHG

Dhlakama to continue patrimonial power grabs and dismiss calls for an actual performance-based platform, the 2008 mayoral elections and subsequent 2009 general elections exposed the structural , and perhaps irreparable, weaknesses of RENAMO.214 At its peak in 1994, RENAMO held 112 seats in the

1DWLRQDO3DUOLDPHQWWR)5(/,02¶VVHDWV 1mRe(VWH23DtV3HOR4XDO1Hgociei E Sonhei"

2009).215 The 2008 municipal elections and very public conflict between Beira Mayor Simango and

$IRQVR'KODNDPDDQGKLVSUR[\0DQXHO%LVVRSRLQ'RQGRIXUWKHUZHDNHQHG5(1$02¶VVWDWXUH

Dhlakama made several statements that severely undermined his credibility, indicating publicly that he spokesperson for the MDM, and a major force in the creation of the MDM, introduced this researcher to a room of 50-60 individuals. He knew the names of all those present and their positions. Three influential bairro secretaries were present.

214 2QHSROLWLFDODQDO\VWLQGLFDWHGWKDW5(1$02LVDSDUW\EXLOWRQ'KODNDPD¶VFXOWRISHUVRQDOLW\ The new elites within the party who are interested in modernization and technical expertise are backed Simango in 2008 while the old combatentes who are used to personalist politics stuck to Dhlakama. Another political analyst mirrored this observation, stating that RENAMO was not a deep opposition SDUW\ZLWKDQDJHQGDIRUGHYHORSPHQWEXWUDWKHU³WKHRQHWKDWVWDQGVXSWR)5(/,02´

215In 2000, Raul Domingos, the lead RENAMO negotiator in the 1992 Rome Peace Accords, was kicked out of the party. Like Simango in 2008, Domingos created another party²the Party for Peace, Democracy and Development- Partido para a Paz, Democracia e Desenvolvimento) and ran for the President in 2004. He lost terribly, having captured only 2 percent of the vote. Simango outperformed him by winning almost 8 percent of the national vote.

208

LVD³GLFWDWRU´ZLWKLQWKHSDUW\DQGWKHIDWHRI0R]DPELFDQGHPRFUDF\OLHVLQKLVKDQGV µ,¶PD

'LFWDWRU¶%RDVWV'KODNDPD,2).216 Dhlakama relied on an anti-FRELIMO strategy instead of offering voters a choice based on what RENAMO would do to effectively govern.217 As a result,

5(1$02VXIIHUHGWKH³PRVWKXPLOLDWLQJORVVRILWVKLVWRU\DVDSDUW\´QRWJDLQLQJDVLQJOHPD\RUDO win or majority in municipal assemblies of all 43 municipalities (Álvaro 200 %\5(1$02¶V

Parliamentary presence shrank by over half to only 51 seats in the National Parliament while FRELIMO gained 60 seats since 1994 for a total of 191 seats (Hanlon 2009).

/DVWO\)5(/,02¶VHIIRUWVWRLQFUHDVHYRWHUDZDUHQHVVOLNHO\LQWended to bolster the dominant

SDUW\¶VSRVLWLRQPD\KDYHEDFNILUHG)URPWRWKH)5(/,02JRYHUQPHQWPDGHYRWHU outreach and enrollment a priority. The intent was to consolidate FRELIMO leadership of municipalities, which it did, increasing the number of FRELIMO municipalities from 28 in 2003 to 42 in 2008. Furthermore, FRELIMO increased the number of politically decentralized municipalities from

33 to 43. Consequently, local elections actually grew in relevance. Despite national trends indicating voter apathy in general and presidential elections, municipal voter turnout has been on the rise (see 6.

3). The first national election in 1994 for president had an 88 percent voter turnout. By 2009, only 45 percent of the electorate voted (see Table 6.3). In contrast, in 1998, only 15 percent of voters cast their ballots in the first municipal elections. By 2008, 46 percent of the electorate voted, a higher turnout than the subsequent 2009 national election. This trend of decreasing turnout in national elections and increasing turnout in local elections appears to be unique in sub-Saharan Africa suggesting that political decentralization may be increasing voter participation. Other cases in sub-Saharan Africa, such as

Uganda and Ghana, national turnout in elections grew in both countries faster than local elections, unlike results in Mozambique (Aworti 2010,629). While in the short term, FRELIMO has consolidated their position this trend of increasing participation in local elections could provide in a platform for other parties to launch a successful bid by national offices.

216 Author interview, Miguel De Brito, EISA, Maputo, March 4, 2009. Author interview, Paul Fauvet, Information Agency of Mozambique (AIM), Maputo, March 4, 2009.

217 Ibid.

209

The real challenge has been (and continues to be) that FRELIMO is the only party with management experience. Afrobarometer 2002 survey results indicated that 61 percent of Mozambicans trusted FRELIMO to just 22 percent who supported RENAMO (Carbone 2005,421). Recent

Afrobarometer findings show that Mozambicans are less likely to experience a shortage of necessities in

2008 than in 2002 or 2005, understandably bolstering confidence in the dominant party (Shenga

2010, )XUWKHUPRUH5(1$02¶VGLVRUJDQL]DWLRQDQGLQFDSDFLW\WRDUWLFXODWHDQ\WKLQJEXWDQDQWL-

FRELIMO platform has limited voter confidence. However, the ability for Simango to build a campaign centered on performance has arguably altered the dynamic. This, in combination with greater voter awareness, may have been the major dynamic at play in the 2009 national elections.

In conclusion, as this chapter has demonstrated, decentralization takes time and resources to implement. One of the town councilors who had been involved in Beira politics for over 20 years related that decentralization came in 1998, passed, and nothing cKDQJHG³2QHQLJKWWKHPD\RUVZHUH mayors of the Executive Council and the next morning they were mayors of the Municipal Council.

7KHIUDPHZRUNFKDQJHGRYHUQLJKWEXWWKHLQVWLWXWLRQGLGQ¶W´218 The results of the 2008 municipal elections and 2009 national elections suggest that municipal performance opens space politically dominated by the opposition party RENAMO to new players like the MDM, which successfully challenged the dominant FRELIMO party in a new and, perhaps permanent, way.

218 Author interview, Town Councilor of Sanitation, Beira, March 12, 2009.

210

Table 6.1: Hypothesis and Findings

Hypothesis: Decentralization creates incentives that motivate municipal government officials to act and opportunities for residents to pressure them.

POLITICAL PRESSURE

Elections:

Expected Finding: Competitive elections incentivize mayors to improve municipal services to constituents.

211

Table 6.2: Expected versus Actual Findings

ELECTIONS M UNICIPAL PERFORM ANCE Expected Actual

MAPUTO

M edium Medium. While the mayor delivered on his campaign High promises to improve municipal management to a broad- based constituency of voters, reforms implemented by the mayor during the 2003 mandate resulted in curtailing benefits to members of the dominant party and cost him the 2008 nomination as the FRELIMO candidate.

BEIRA

High +LJK7KHPD\RU¶VSHUIRUPDQFHLQGHOLYHULQJRQ High campaign promises combined with weakness of the opposition party led to a groundswell of support, which led to a successful independent bid for Mayor in 2008 and the eventual creation of a new national party.

DONDO

Low Low. While the pressure was high from the party on the High mayor to perform, voters consistently re-elected the mayor and so the political from them was low. The PD\RUKDVEHHQ'RQGR¶VRQO\PD\RUVLQFHWKHILUVW municipal election in 1999 and has been regarded as an effective mayor, consistently supported by the dominant party.

212

Table 6.3: Electoral Turnout

M unicipal National

1994 * 88%

1998/9 15% 70%

2003/4 24% 36%

2008/9 46% 45%

Source: Mozambican Constitutional Court; Carbon 2005: 420; Mozambique Peace Process Bulletin Issue 21, July 21, 1998; Mozambique Political Process Bulletin, 2008 Local Election Issue 21, December 9, 2008; EISA 2009 MZ Presidential Elections website, updated January 2010. *Municipal elections were first held in 1998. National elections are held the year after municipal elections.

213

Table 6.4. 2003 Municipal Election

M unicipality Candidate Party Registered Turnout % Turnout Maputo Eneas Comiche FRELIMO 96,035 75.15% Phillipe A. Gagnaux JPC 12,695 9.93% Carlos Jeque PDD 2,543 1.99% Artur Vilanculos RENAMO 15,289 11.97% Pedro Loforte PT 1,223 0.96% Total, Maputo: 605,529 128,871 21.28%

Beira Davis Simango RENAMO 29,610 53.43% Djalma Lourenco FRELIMO 23,378 42.23% Antonio Romao IPADE 1,451 2.62% Pedro Langa PIMO 956 1.72%

Total, Beira: 215,326 57,553 26.73%

Dondo Manuel Cambezo FRELIMO 6,641 61.61% Joaquim Greva RENAMO 3,478 32.27% Matias Traquino IPADE 660 6.12%

Total, Dondo: 38,635 11,417 29.55% Notes: Total turnout does not add to total because votes include all votes cast, including invalid votes cast but later discarded. Constitutional Council 2003 Deliberation no. 16/CC/04 January 14, 2004; Deliberation no. 5/CC/05 January 19, 2005; Accord No. 02/CC/2009, Annex 1 and 2.

214

Table 6.5. 2008 Municipal Elections

M unicipality Candidate Party Registered Turnout % Turnout Maputo David Simango FRELIMO 255,050 85.78% Antonio Namburete RENAMO 42,281 14.22%

Total, Maputo: 661,034 308,323 46.64%

Beira Deviz Simango Independent 76,734 61.61% Lourenço Bulha FRELIMO 42,003 33.73% Antonio Romao PDD 1,718 1.38% Filipe Alfredo PIMO 768 0.62% Manuel Pereira RENAMO 3,316 2.66% Total, Beira: 230,070 129,628 56.18%

Dondo Manuel Cambezo FRELIMO 12,203 76.75% Manuel Bissopo RENAMO 3,053 19.18% Carlos Jeque PDD 658 4.17% Total, Dondo: 33,492 17,340 51.77%

Notes: Total turnout does not add to total because votes include all votes cast, including invalid votes cast but later discarded. Constitutional Council, Accord No. 02/CC/2009, Annex 1 and 2.

215

VII. CONCLUSION: INSIGHTS AND PATHS OF INQUIRY

Institutions do not merely reflect the preferences and power of the units constituting them; the institutions themselves shape those preferences and that power.

Robert Keohane (Keohane 1998, 382)

«WKHQHZLQVWLWXWLRQVKDYHimproved service quality for the people. Despite their political differences, all local actors strongly agreed upon the point that with the local autonomy of Autarquias, service delivery had become better, quicker, more GHSHQGDEOHDQGPRUHWRWKHSHRSOH¶VQHHGVWKDQLWKad been under the former regime RIGLVWULFWDGPLQLVWUDWLRQ«7KXVLQWKH0R]DPELTXHH[SHULHQFHGHFHQWUDOL]DWLRQ PHHWVLWVSURPLVH³to bring the state closer to the people´

Wolf Linder (Linder 2009, 25)

Introduction

.HRKDQH¶VDQG/LQGHU¶VTXRWHVDERYHLOOXVWUDWHKRZLQVWLWXWLRQDOFKDQJHVKDSHVSUHIHUHQFHV and power both in theory and in practice. I set out in this dissertation to examine whether revenue and political pressures compelled municipalities to deliver better services to their residents. Specifically, I analyzed how three factors²namely participation, revenue and elections²affected municipal performance. I hypothesized that decentralization created incentives that motivated government officials to act and opportunities for residents to pressure them (see Table 7.1). As new authorities put resources in the hands of local decision-makers and votes (and greater voice) in the hands of citizens, resource pressure and revenue pressures over these municipal officials motivates them to improve municipal performance to gain access to increased revenue and votes. I asked why variation in municipal performance exists within cases in which decentralization has occurred and institutional factors such as historical context, elections and revenue authorities remained constant, as was the case

ZLWK0R]DPELTXH¶VGHFHQWUDOL]DWLRQ6HOHFWLQJWZRPXQLFLSDOVHUYLFHVLQHDFKRIIRXU municipalities within Mozambique, I explored my hypothesis, expecting in a post-conflict, post- socialist, low-income new democracy to find weak citizen participation that cannot pressure the

216

government to improve performance, municipal governments with little control over resources, a high degree of dependence on donors and a central government and figure-head mayors subject to little voter scrutiny and political competition in a one-party state.

I found that, even in the context of weak participation, low resource capacity, and weak political competition, municipal performance was high in three cases due to the following reasons.

First, I found that, while no single civil society organization single-handedly improved municipal performance, participation and voice exercised a role in changing and facilitating municipal policy in three of the four cases in favor of improved municipal service in the two test areas I studied; namely, trash collection and market management. Second, unearned income improved municipal performance and own-source revenue generation. Third, mayoral leadership was critical in improving municipal services. I also found that political decentralization opened up space for performance to be a major driver in helping a high-performing opposition mayor compete and win his second term and create a new national party that captured seats in the National Parliament. While these three findings existed in different combinations across civil society, revenue and elections for the three high-performing municipalities, the fourth case where I found performance to be low had weak civil society, majority unearned income, and low political competition in elections (see 7.2).

In Chapters 1 and 2, I contended that, while decentralization had been embraced by donors as a policy prescription thought to strengthen fledgling democratic governments in weak states by promoting accountability and service provision, that seemingly little attention was paid to what factors at the micro-level influence municipal performance. In these chapters, I explored theories that explained observations of how revenue and political pressures compel government accountability and performance. After testing for variation in my dependent variable in Chapter 3, I spent the remaining three chapters analyzing how my evidence responds to my original hypothesis and contributes to the theoretical debates related to the effects participation, revenue and elections have on municipal performance in Maputo, Matola, Dondo and Beira. In this final chapter, I will revisit my hypothesis and reiterate and extend my conclusions related to the relationship between civil society, revenue, and elections to municipal performance. 217

The Hypothesis, Revisited

I hypothesized that decentralization improves municipal performance by promoting accountability.

Specifically, decentralization creates incentives, which motivate municipal government officials to act and opportunities for residents to pressure them in the following ways:

1. Resource pressure over local officials increases as new authorities put resources under the

control of local decision-makers; and,

2. Political pressure over local officials increases through new opportunities for the expression of

citizen preferences..

(or reward) municipal performance by withholding votes and revenue. Furthermore, by subjecting local officials to the pressures of earning both votes and revenue, citizen participation should increase as civil society also develops leverage because it can, in theory, now galvanize voters and tax payers to sanction or reward public officials. As fiscal decentralization scholar Roy Bahl summarizes, decentralization

³JHWVJRYHUQPHQWFORVHUWRWKHSHRSOH´EHFDXVHLWVKLIWVGHFLVLRQ-making and authorities to the local

OHYHOHQDEOLQJYRWHUVZKRSD\WD[HVWR³IHHOWKDWWKH\FDQKROGORFDORIILFLDOVDFFRXQWDEOHIRUGHOLYHU\ of services at some acceptable quantity and quality´ %DKO 1999, 60-61).

,IRXQGWKDWUHVRXUFHSUHVVXUHDVH[HUWHGE\DIHHOREE\H[LVWHGLQ0DSXWR¶VLQIRUPDOYHQGRUV

DVVRFLDWLRQ,DOVRVDZWKDWGHFHQWUDOL]DWLRQ¶VQHZWD[DWLRQDXWKRULWLHVHQDEOHGDOOIRXUPXQLFLSDOLWLHVWR increase own-source revenue in the presence of large shares of unearned income from 2004 to 2007.

218

However, these increases alone did not independently improve performance. Donor projects targeted to increasing capacity to design fee schedules, collect fees and improve municipal performance helped municipalities improve service in three of the four cases in the two test areas. Political pressure also motivated improvements in municipal service. Strong and active civil society organizations facilitated the provision of municipal services in Beira, Dondo, and Maputo. Citizen protest in Maputo resulted in a more rational and acceptable fee schedule. Furthermore, in the case of Maputo donors helped neighborhood organizations pressure the CMM to gain access to contracts. Participatory budgeting in

Dondo provided citizens a channel to municipal decision-making. In Dondo, Maputo, and Beira,

Merchant Associations were engage interlocutors between vendors and these municipalities.

Competitive mayoral elections in which the candidates could vie for votes based on fulfillment of campaign promises occurred in two of the four cases and, in three of the four cases, mayors were technocrats and experienced managers who could assemble and motivate their management teams to improve municipal performance.

To summarize, I found that in each of the three cases where I found performance to be high, at least one of the three factors was present. In Maputo and Beira I found three factors. Maputo had a strong civil society, a majority share of earned income and a moderate level of political competition.

Beira had an active civil society, a majority share of earned income, and a high level of political competition. In Dondo, I found one factor present: an active civil society. In Matola, I found none present.

Civil Society and Municipal Performance: Strategic Influence

I hypothesized that strong civil society would pressure a municipality to improve performance in the areas of trash collection and market services. I measured the strength of civil society by examining two factors: its access and control to resources and influence over municipal policy. I found that strong civil society in and of itself was not a pre-requisite for strong municipal performance and that civil society changed over time. I found that citizen participation was present and influenced 219

municipal performance in the two test areas in three of the high performing cases--Maputo, Beira, and

Dondo--to varying degrees. I found that the type of donor assistance and relationship with the municipality were greater determinants of how participation positively affected municipal performance than its strength. While a strong civil society was not a prerequisite for high municipal performance, I found that the three cases in which performance was high had civil society organizations that facilitated and/or extended the municipalities reach. Maputo was the only case in which I found civil society in the form of ASSOSTI and start-up neighborhood organizations act as an independent political force in lobbying the CMM for a share of resources. It is also the only case in which citizen protests over trash fees actually led to two positive outcomes improving trash collection. First, it caused the CMM to revise its fee schedule so it shifted the costs to those who produced more trash and could afford to pay more, resulting in a more sustainable and more equitable fee schedule. Second, it forced the CMM to work with neighborhood organizations, that were granted access to municipal contracts, became self- sustaining microenterprises, which improved the delivery of municipal trash collection.

I IRXQGDVWURQJOREE\WKDWZDVDEOHWRIRFXVWKH&00¶VDWWHQWLRQRQLQIRUPDOYHQGRUVEXWWhis attention and change in municipal policies in response to this lobby did not lead to a direct improvement in municipal services. $66267,¶VLQGHSHQGHQWSROLWLFDOIRUFHDV a strong lobby was motivated more by its own interests than those of its members and civil society extended municipal services. In Beira,

Merchant Associations acted as facilitators in delivering CMB services. In Dondo, a donor-funded participatory governance process allowed residents to influence the types of infrastructure projects the municipality funded. In Beira, Dondo and Maputo, donors played a a role in facilitating and financing these changes.

As I discussed in Chapters 2 and 4, I examined how in a post-socialist new democracy such as

Mozambique, civil society organizations outside the state lobby for a greater share of state resources as was the case with ASSOSTI and some neighborhood trash organizations while organizations within the state extend both the state¶s reach and ability to provide services, as was the case with Merchant

Associations. As I discussed in Chapter 4, I found that Maputo had strong civil society organization

220

whereas Beira and Dondo had active organizations that did not challenge municipal authority. I also found that participation in a the form of a boycott in Maputo of trash fees and a participatory budget process in Dondo have citizens a voice in municipal decision-making. Lastly, I found participation to be weak in Matola. While strong civil society organizations were not a prerequisite, my evidence in

Maputo, Beira and Dondo suggested that at least two of three factors were present: financial sustainability, ability to influence municipal policy, and capacity-building donor assistance. In

Maputo, I found all three. In Dondo, I found capacity-building donor assistance and the ability to influence municipal policy whereas in Beira, only the latter was present. In Matola, none of these factors was present.

My research indicated that donors supported building the capacity of civil society organizations that contributed to improvement in municipal performance (although to varying degrees) in three of the four cases. ,Q'RQGRDQG0DSXWR DQGOHVVHUH[WHQWLQ%HLUD¶V0XQKDYDVDQLWDWLRQ project), donor aid built capacity and municipalities partnered with them to deliver services. In

Maputo, boycotts of trash fees caused the municipality to revise its fee schedule, with the technical assistance of AGRESU into a schedule which is now a national model and (more importantly) was a level fee-payers were much more willing to pay. This, in turn, institutes a self-VXVWDLQDELOLW\WR0DSXWR¶V trash collection services that was absent before. Trash organizations also benefitted from donor assistance in pressuring the CMM into providing them contracts to collect trash in informal areas. In

'RQGRWKH$XVWULDQ'HYHORSPHQW&RUSRUDWLRQIXQGHG'RQGR¶VLQLWLDOH[SHULPHQWZLWKSDUWLFLSDWRU\ budgeting, which continues to be a process by which the CMD gathers input on infrastructure allocations annually.

In Beira, Africa 70 worked closely with community leaders and organizations in Munhava

QHLJKERUKRRGWRLQVWDOODQGXSJUDGHVHFWLRQVRI%HLUD¶VZDWHUVDQLWDWLRQV\VWHP,QDOOWKUHHFDVHV

Merchant Associations acted as interlocutors between the municipality and vendors to improve infrastructure and sanitation in the markets. However, only in Maputo were these associations able to gain control and access to a portion of fees collected to improve markets as they saw it. In Matola, these

221

conditions were absent. Evidence pointed to low-level corruption both in the provision of trash collection services and the selective deposits of market fees by market chiefs. Merchant Associations were weak and neighborhood trash associations absent..

However, even in the three cases where participation and voice contributed to improved municipal services, there are examples of how participation did not have a positive effect on municipal performance. In Maputo, ASSOSTI was a strong, politically-connected lobby with access to membership fees to sustain its operations. While this informal vendor lobby did influence municipal policy, it did so to primarily secure access to membership fees. In Beira, Munhava neighborhood trash organizations were active only when donor funds were present. In Inhamodima neighborhood, the CUB project actually ended poorly due to miscommunication on remuneration. Even some donor interventions were not successful in these same contexts. In Dondo, while the participatory budget process connected residents to the municipality and provided a vehicle to inform infrastructure investment allocations, ultimately, local scholars concluded that it was unclear how much the CMD relied on that feedback in making those allocations.

The 2002 protests in Maputo over the introduction of the flat trash fee at the peak of inadequate municipal collection provided another example of how participation and voice caused a change in municipal policy. It is also the only case in which citizen protests over the introduction of the trash fee actually pressured the CMM to withdraw these fees, do outreach, and reintroduce another approach that was more consistent with citizen expectations. By 2003, a fee schedule had been adopted and implemented in Maputo that was met with a better reception by its citizens. Throughout the process,

AGRESU worked closely as a technical advisor to the CMM in introducing and then re-introducing the revised fee scale and working to create a sustainable, multi-year plan for financing trash collection in

Maputo through fee revenue. Furthermore, this opening for citizens to negotiate with the CMM also opened the CMM to working with neighborhood organizations as contractors to extend trash collection to notoriously inaccessible informal neighborhoods

222

In terms of market management, Maputo was also the only municipality, in which the CMM was able to assert its authority to collect vendor fees by allowing Merchant Associations more say over how to manage their markets with 10 percent of vendor fee revenue because the CMM needed assistance in managing vendors and collecting fees. Furthermore, as evidenced by the three episodes in which ASSOSTI challenged the CMM, this lobby boycotted payments in order achieve municipal policies favorable to itself. Where the state, in this case the CMM, was absent, the community filled in the gap and compelled the CMM to adapt its policies in favor of them²either through influencing how fees were collected, as in the case of ASSOSTI, or pressuring for contracts. The World Bank,

AGRESU, CARE and Médecins Sans Frontières were all donors that provided capacity-building technical assistance to civil society organizations that helped them apply political pressure by evolving into viable contractors that could work with the CMM to improve service.

To reiterate, Beira and Dondo had active civil societies, and in Dondo, donors played a prominent role in improving service in the two test areas. In Beira, Mayor Simango and his designates instituted visible improvements to trash collection and vendor services. Merchant Associations in the largest markets worked closely with the CMB and, in many cases, were appointed at the start of the

Simango Administration in 2003 so they had explicit ties with the new leadership. Major market improvements in Maquinino caused shifts between formal and informal markets that occurred peacefully, with the addition of 200 additional stalls, two roofs, water pumps and drainage lines.

Fountains were also installed by the CMB in Goto, the largest informal market. Focus groups report immediate improvements in trash collection within months of Simango taking office in 2005. While donors and other external actors like ESSOR, Africa 70, and the Catholic University of Beira were present in Beira, in the area of trash collection, they were unable to provide capacity building to neighborhood organizations to foster their self-sustainability.

In Dondo, the Austrian Development Corporation funded the participatory budget process in place today in which hearings provide input directly to CMD town councilors for municipal infrastructure investments that include market infrastructure and trash collection equipment. Merchant

223

Associations in decentralized markets collect fees on behalf of the municipality. While civil society organizations themselves are not strong, Merchant Associations acted as proxies for the municipality and the participatory budget process provided feedback mechanisms for both neighborhoods and the

CMD in limited, but important, ways. In Matola, the absence of neighborhood trash organizations, weakness of Merchant Associations in the major markets, and presence of politically-motivated favoritism in the distribution of public goods combined to detract from the provision of high-quality trash collection and market management.

Revenue and Municipal Performance: Targeted Use of Unearned Income to Support Growth in Earned

Income

In Chapters 2 and 5, I presented evidence that showed improvement in both municipal performance in three cases and explosive growth in own-source revenue across four cases from 2004 to

2007. The literature underplayed the positive role unearned income²such as targeted aid and central government transfers²can play. However, evidence of research done on municipal revenue (this study included) suggests that, even in the context of high aid dependence on central government transfers ,, municipalities are improving performance. One study of aid on the quality of governance of two

Mozambican municipalities concluded that the municipality with more aid had a higher quality of governance (Nuvunga 2007, 54-71). A World Bank study of 18 sub-Saharan African countries suggests that for every dollar in foreign aid, USD$0.89 goes to government expenditures and a relatively small share (USD$0.11) goes to tax relief, thereby directly impacting government performance by increasing expenditures without drastically decreasing taxation incentives (Devarajan, Rajkumar, Swaroop 1999).

In Tanzania, donors empowered local bureaucrats to charge taxes in the face of opposition from elected officials and the public (Fjeldstad 2001, 301). Even within countries where aid dependence is extremely high and the prospects of raising revenue at the central government level seem limited, aid given to local governments positively affecting the quality of government.

224

7KLVVWXG\¶Vfindings support the case that unearned income did not dissuade the pursuit by these municipalities of own-source revenue with new taxation authorities even in the presence of majority shares of unearned income. This finding builds on the work Paul Collier has done in analyzing how different types of revenue²QDPHO\ZKDWKHWHUPV³VRYHUHLJQUHQW´VXFKDVRLOUHYHQXHVDQG

³VFUXWLQL]HGUHYHQXHV´VXFKDVDLGDIIHFWVJRYHUQPHQWSHUIRUPDQFH(Collier 2006; Collier and Dehn

2001). The dynamic of earned and unearned income I explored in Chapter 5 reflects &ROOLHU¶V

FRQFOXVLRQWKDW³DLGFRPHVZLWKYDULRXVGRQRU-imposed mechanisms of scrutiny, which may spill over into other expenditures, and so substitute for reduced pressure from citizens´ &ROOLHU

Maputo and Dondo are two examples of such a dynamic.

$LG³MXPSVWDUWHG´SHUIRUPDQFHDQGIHHFROOHFWLRQLQWKHWHVWDUHDVLQthese two municipalities.

Technical assistance was one of the accountability factors that Collier discussed that incentivizes the delivery of public goods as a result of scrutinized revenue (namely, aid) in the absence of public participation. In both Maputo and Dondo, donors provided specific, service-related technical assistance.

PROMAPUTO , a 10-year, municipality-driven community development and urban revitalization plan that DLPVWRIXQGSURMHFWVWKDWVXSSRUWWKH&00¶VORQJ-term development goals in a comprehensive fashion, is such an example. Updating municipal regulations, to include revised fee schedules, improving market infrastructure and trash collection equipment, updating computers in City Hall are all

WKHW\SHVRIDFWLYLWLHVIXQGHGXQGHU3520$3782$*5(68¶VWHFKQLFDODVVLVWDQFHIRFXVHG specifically on determining the fee schedule for trash collection and the feasibility studies to back up both the fee levels based on electricity consumption and the multiyear plan to increase fees and reduce reliance on external funding for trash collection. In Dondo, the Austrian Development Cooperation

IXQGHGLWVSDUWLFLSDWRU\EXGJHWSURFHVVVWDUWLQJLQZKLFKFRQWLQXHVWRLQIRUPWKH&0'¶VGHFLVLRQ- making on investments of its infrastructure budget. Both examples incentivized and enabled the delivery of public goods by providing accountability feedback mechanisms that frame resource allocation decisions in both municipalities.

225

I also observed both a shift in the character of central government transfers and an explosion of own-source revenue which suggests that municipal performance has benefitted from not only the availability of more resources but also a diversification of those resources favorable to improved performance. Central government transfers shifted from municipal personnel compensation to infrastructure investment in 2004 to 2007 across all four cases. On average, municipal personnel compensation, the largest transfer from the central government, decreased by 20 percent as a share of overall transfers from 65 percent in 2004 to 45 percent in 2007 while infrastructure transfers increased 7 percent from 31 percent in 2004 to 38 percent in 2007. This shift implies an increased relative investment in infrastructure rather than subsidizing personnel. Furthermore, municipal personnel compensation shrunk as a share of overall receipts which points to an increase in municipal staff independence since their salaries rely less on central government transfers and more on other sources of municipal income. Own-source revenue almost doubled even though, on average, 41 percent of revenue from 2004 to 2007 came from central government transfers. Even as central government transfers exploded 164 percent from 2004 to 2007, own-source revenue roughly kept pace with a 135 percent in this same period. This suggests that case study municipalities were not deterred from using new taxation authorities in the presence of large (and growing) shares of central government transfers.

It further indicates that, overall, these four municipalities experienced a massive increase in resources.

While an increase in resources alone does not guarantee better service delivery, in three of the four cases, these resources were coordinated and put to work as evidenced by observations of the two test services.

As stated above, my findings suggest that own-source revenue growth was not disincentivized by the presence of large shares of unearned income. Additionally, unearned income²specifically, aid²improved municipal performance and own-source revenue collection. A further area of study specifically in Mozambique and in states with a high reliance on unearned income is the relationship between budget support and performance of decentralized subnational units, like municipalities. While the four municipal budgets I studied contained on average only 12 percent aid to overall revenue, a low

226

level for a country in which at least 50% of public spending between 1994 to 2004 at the national level was derived from aid, 30 percent of the total development assistance is delivered through the regular budget (Batley, Bjørnestad, and Cumbi 2006, S7; Saide and Nhata 2007, 76). As a result, while explicit aid to municipalities may be low as identified in their budgets, central government transfers can continue to contain large shares of aid over which localities have no control. This could give central governments more ability to punish municipalities by withholding transfers. It can, alternatively, exert the kind of accountability pressure on the government DV³VFUXWLQL]HGUHYHQXHV´DV&ROOLHUVXJJHVWV

Either way, as more countries decentralize and donors opt to support country systems, additional study of how this modality of assistance affects performance incentives at both the national and local levels is warranted.

Another area in which to extend this inquiry would be how aid to local governments affects political accountability at the local level. Collier noted the link between scrutinized revenue (aid, in this case) and winning electoral strategies based on performance. In his study of aid to African states,

Collier found that donors can and do apply both revenue and political pressure on the state to compel more efficient and effective service provision because donors hold governments accountable. He noted that aid had positive effects on the performance of African governments and economies in comparison to that of resource rents. Collier further noted a link between scrutinized revenues, like aid, and electoral campaign strategies based on service delivery because the underlying dynamic is that donors hold states accountable. In fact, Collier theorized that

«WKHODUJHU«VFUXWLQL]HGUHYHQXHV DUH WKHPRUHSURPLQHQW«SXEOLFJRRGV provision in political debate. Hence, as sovereign rents are diminished and scrutinized revenues increased, there comes a point at which, even in social conditions of ethnic loyalties and poor information, the winning electoral strategy becomes effective delivery of public goods rather than patronage (bold added, Collier 2006, 1484).

I found empirical evidence of this dynamic in Maputo, Dondo and Beira since all three mayors relied on performance improvements in their campaigns, which were related to donor interventions to improve municipal service. Moreover, I found in Matola that the lack of municipal capacity and leadership 227

dissuaded additional donors--as evidenced by the absence of PDM grant investments--from working in donor demonstrates the seminal role that leadership plays in gaining access to resources to improve municipal performance. In the 2003 and 2008 municipal election campaigns, Mayor Comiche in

Maputo, Mayor Cambezo in Dondo and Mayor Simango in Beira articulated performance-based platforms. While to a lesser extent in Beira, in three cases, donors²whether through the World Bank and GTZ/AGRESU in Maputo, Austrian Development Corporation in Dondo or the Italian Cooperation in Beira²funded projects that improved trash collection and market management so these mayors could make the case that they delivered on campaign promises to improve municipal services. Beyond providing technical assistance and infrastructure funding, aid also provided political incentives to compel municipal performance. In the following section, I discuss this relationship between donors, improvements in municipal performance and the 2003 and 2008 municipal elections further and how political pressure in the context of these two races not only affected municipal performance but contracted and expanded political spaces.

Elections and Municipal Performance: Contracting and Expanding Spaces

In the penultimate chapter, I presented my evidence of how elections affect municipal performance by examining how the mayors in Maputo, Dondo and Beira fulfilled campaign promises related to trash collection and market management in the period bookended by the 2003 and 2008 municipal elections.

I hypothesized that competitive elections incentivize mayors to improve municipal services to constituents in fulfillment of campaign promises because mayors seek re-election (Table 6.2). First, I found that dominant party politics affected the ability for voters to exert political pressure through votes for a dominant-party mayor. Second, mayoral performance and constituent votes mattered most for re- election in the opposition municipality. Lastly, I found that mayoral leadership was the driving factor behind municipal performance. Leaders initiated and facilitated changes to the way the municipality worked, enhanced trust in key stakeholders, such as donors, and engaged with citizens in such a way as

228

to reveal improved municipal services in three of the four cases I studied. I discuss this last component in the next section.

New democracies tend to have weak parties and strong states so once in power, dominant parties will generally co-opt state functions to preserve power. This disincentivizes performance of locally- elected politicians because their reward is not re-election as expressed through voter preferences but rather service to the party, thereby discouraging public performance and service provision in favor of capturing benefits for the dominant party. Although elections can and do serve as the sanction, the sanction is far less important in the dominant party because their primary purpose is to capture benefits not curry public support. I found this dynamic DWSOD\LQ0R]DPELTXH¶VDQGPXQLFLSDO elections. FRELIMO mayors won all 33 municipalities in 1998, 28 municipalities in 2003 and 42 of 43 municipalities in 2008. This evidence demonstrates the dominance of the FRELIMO party and limitation of voter choice. Furthermore, it reveals what some might observe to be the recentralization of

FRELIMO political power because the main opposition party²RENAMO²failed to win any municipalities in 2008 and in the 2009 national elections, the party leader and Presidential candidate²

Afonso Dhlakama²garnered the lowest share of votes since the first Presidential elections were held.

The 2008 and 2009 elections showed increasingly that RENAMO as a viable alternative and challenger to FRELIMO was severely, and perhaps irreparably, weakened which calls into question the long-term viability of Mozambican democracy.

I found some trends in my own study of four municipalities that reflect these observations.

FRELIMO made strategic choices about Maputo, Dondo, and Beira candidates. In Maputo, FRELIMO choose not to re-nominate Dr. Comiche in 2008 because he revealed himself WREH³too independent´.

The fact that Dr. Comiche did not try to mount an independent candidacy as did Mayor Simango in

Beira reflects a dynamic at play that is not unique to Mozambique but reflective of the dominant party context. Dr. Comiche was perceived to be more of a technical expert than a politician with a grassroots base and, while modernizing the CMM, made decisions that politically cost him. When he was not re- nominated by the party, he did not mount an independent candidacy as did Simango in Beira. Some

229

observed that it was because he lacked the grassroots base, Comiche himself was a party loyalist, and that the party was too strong to allow that type of a challenge had Comiche presented it.

In Dondo, FRELIMO continued to back a high-performing mayor in Manuel Cambezo to counterbalance the effects of Mayor Simango in Beira, in Sofala province, a RENAMO-leaning province and home to Afonso Dhlakama. In Beira, FRELIMO put up a candidate²Laurenço Bulha² familiar to voters because he had both run in previous elections and was the major backer of the largest shopping center in downtown Beira. FRELIMO invested a great deal in advertising for this candidate in comparison to Simango with the intent of capturing Beira back. However, this candidate lost. While this outcome reflected a huge victory for the independent candidacy of Simango, it reflected the hollowness of RENAMO leadership. The RENAMO candidate in Beira had an even poorer showing than the FRELIMO candidate, reflecting the political mistake RENAMO made in jettisoning Simango.

This development reflects a weakening of political competition and a contraction of political space in Mozambique because the major opposition party has reached its lowest level of support in both mayoral seats and in votes at the national level since 1994. I found, however, two interesting dynamics at play. First, in two of the three elections, FRELIMO nominated candidates based performance.

Second, I found that mayoral performance mattered more in the one municipality that FRELIMO did

QRWZLQLQEHFDXVHRIZHDNQHVVRIWKHRSSRVLWLRQSDUW\WKHPD\RU¶VDELOLW\WRIXOILOOFDmpaign promises, grassroots popularity, and willingness to challenge both the dominant and major opposition parties.

FRELIMO was able to consolidate its power prior to the first municipal elections in 1998 by arguably selecting the 33 new, politically decentralized municipalities without RENAMO and then running essentially uncontested because RENAMO boycotted the elections, enabling FRELIMO to win all 33 mayoral seats. However, by the 2003 elections, FRELIMO jettisoned 20 of its mayors, including

Maputo Mayor Cananá and re-nominated only 13 incumbents who ran on performance (Nuvunga 2004).

)5(/,02¶V0DSXWRQRPLQHH'U&RPLFKHGLGQRWKDYHPXQLFLSDOH[SHULHQFH+RZHYHUWKH

230

IORRGVDQGSUHYLRXVPD\RU¶VFRUUXSWLRQVFDQGDOVFUHDWHGDQRSHQLQJIRr this mayor to run a campaign based on improving trash collection and services to vendors. While some contended that

Mayor ComicKH¶V$GPLQLVWUDWLRQIDYRUHGWKHZHDOWKLHUVHFWLRQVRIWKHFLW\0D\RU&RPLFKH consistently sought to professionalize the CMM, instituting policies for review of contracts that did not favor FRELIMO. He instituted the 10-\HDUUHGHYHORSPHQWVWUDWHJ\WKURXJKWKH:RUOG%DQN¶V

PROMAPUTO program, trash and vendor fees increased, roads were paved and streetlights installed.

Vendors I interviewed thought he was treated unfairly in not being re-nominated because he did a

JRRGMREDQGPHPEHUVRIFLYLOVRFLHW\VWDWHGWKDWKHZDVVHHQDV³LQFRUUXSWLEOH´0D\RU&RPLFKHZDV not re-QRPLQDWHGLQSDUWEHFDXVHKHZDVVHHQDV³WRRLQGHSHQGHQW´DQGWKH)5(/,02SDUW\VRXJKWWR back a more conformist candidate in David Simango in 2008. Some individuals I interviewed speculated that Mayor Comiche was a party loyalist who would not challenge this decision, suggesting that the FRELIMO party controlOHGWRRPXFKDQG&RPLFKH¶VJUDVVURRWVEDVHZDVWRRZHDNWRPRXQWDQ alternative candidacy. AFROBAROMETER results show that Mozambicans indeed see FRELIMO as the party that can manage, suggesting that it is the only party with experience and really the only alternative to improving the economic well-being of Mozambicans, a sentiment I heard echoed in conversations. While this may have been true, politically decentralized authorities created the pressure for FRELIMO to have to compete on a performance-based platform in 2003 and for Mayor Comiche to have been nominated against a scandal-ridden mayor, a condition that did not previously exist. This is consistent with recent findings related to Mozambican decentralization. Linder found that, while democratic cRQVROLGDWLRQDWWKHPXQLFLSDOOHYHOZDV³PL[HG´³GHFHQWUDOL]DWLRQSURYLGHGFKDQJHVIRU political change currently non-H[LVWHQWDWWKHQDWLRQDOOHYHO´ /LQGHU, 8QGHU&RPLFKH¶V

PDQGDWHPXQLFLSDOVHUYLFHVLPSURYHGDQG0D\RU&RPLFKH¶VSHUIRUPDQFH while not garnering him party support, was acknowledged in the public discourse.

As I observed in the 2008 election in Beira, Mayor Simango, one of five RENAMO mayors elected

LQZDVOHVVKDPVWUXQJWKDQ0DSXWR¶V0D\RU&RPLFKHEHFDXVH5(1$02¶VFUHGLbility and access

WRUHVRXUFHVZDVQRWDVVWURQJDV)5(/,02¶V6LPDQJRKDGDVWURQJJUDVVURRWVSRZHUEDVHLQ%HLUD

231

that was loyal to him, not the RENAMO party and so he, unlike Comiche, was able to mount a successful independent candidacy. Of the four cases, the only case in which I was able to see elections work as a reward was in Beira. Focus groups in both formal and informal communities that I conducted just a few months after the elections indicated that Simango was elected in 2008 because of his visible performance in cleaning up the trash in what was one as one of the sickest cities in the country. As a result, Simango was able to rely on public support to open up democratic spaces and therefore, earned votes by providing public services. So, although the opposition party may be what Barry Weingast calls

D³QDWLRQDOORVHU´DVFDQEHREVHUYHGIRU5(1$02E\UHSHDWHGQDWLRQDODQGORFDOHOHFWLRQVUHVXOWV decentralization provides non-dominant party candidates multiple opportunities to build political momentum in one-party states. Barry Weingast described the dynamic best:

In a decentralized state, however, losers can typically maintain a local power base from which to remain politically visible and to provide benefits to their constituents. Local political strongholds also provide a base from which this party can launch a future attempt to recapture national power. By lowering the stakes of power, decentralization makes it more likely that losers of national elections will give up power (Weingast 2009, 279-292).219

:HLQJDVW¶VSURSRVLWLRQDSSHDUHGLQSUDFWLFHLQP\ILQGLQJVLQ%HLUD,QWKHHOHFWLRQV6LPDQJRUDQ

RQFOHDQLQJXSWKHFLW\DQGLQKLVFDPSDLJQVORJDQZDV³+HSURPLVHVDQGGHOLYHUV´,QEHWZHHQ trash collection improved almost immediately and remained consistent, cholera rates decreased, fee collections grew, and market stalls, water fountains and bathrooms were constructed. Simango built a grassroots powerbase in informal neighborhoods that supported his independent candidacy so much so that the momentum, which started in the mayoral elections of 2008 and grew into a new party²the

MDM. The MDM launched a national campaign and, while some Mozambican analysts indicate that the

219 Carrie Manning highlighted the applicability of this insight in the dominant Frelimo party within the 0R]DPELFDQGHFHQWUDOL]DWLRQFRQWH[WVWDWLQJ³WKHFUHDWLRQRIDXWRQRPRXVPXQLFLSDOLWLHVKDVFUHDWHG additional political arenas that provide potential venues in which differentiation within the party might bHRSHUDWLRQDOL]HG´³$VVHVVLQJ$GDSWDWLRQWR'HPRFUDWLF3ROLWLFVLQ0R]DPELTXH7KH&DVHRI )UHOLPR´7KH)DWHRI$IULFD¶V'HPRFUDWLF([SHULPHQW, ed. Leonardo A. Villalon and Peter Von Doepp, Bloomington and Indianapolis: Indiana University Press, 2005: 236.

232

party is too new and inexperienced to be a credible threat to FRELIMO, the MDM won three seats in the National Parliament in 2009.

My research reiterates the need for further empirical study of the microlevel effects of political decentralization on the national politics of dominant party states. While FRELIMO continues to dominate the Mozambican political landscape, space is being created through local elections as evidenced by the MDM and, while only time can tell whether this space will expand or contract, political decentralization in Mozambique has provided an opportunity that Beira residents in 2008 and

Mozambicans in 2009 used. Political pressure in the form of elections also elevated municipal performance as a necessary consideration of the suitability of candidates for public consideration for both parties as evidenced in the 2003 FRELIMO nominations and the 2008 elections, as discussed.

Further empirical study of these dynamics would paint a broader picture of the direction of political decentralization and its ability to institute consistent and lasting improvements in government performance.

Leadership and Citizenship on Municipal Performance: Factors to Consider

Lastly, as I discussed in the previous chapter, I found that leadership was the intervening variable that managed resources and relationships both within and outside of the municipality that contributed to improvement in performance from the 2003 to 2008 municipal elections in three municipalities. In all four cases, budgets increased from 2004 to 2007, making available more resources to invest in trash collection and market management. All four municipalities had elections and three of the four had active or strong civil society participation in the two services studied. However, in the three cases with high municipal performance (Maputo, Beira, and Dondo), I found mayors who appointed technically competent managers, created direct channels of access between staff and the Mayor and the Mayor and the public to facilitate accountability, and had relationships with donors that provided project support and technical assistance. 233

,KDGK\SRWKHVL]HGWKDWHOHFWLRQVZRXOGLQFHQWLYL]HPD\RU¶VWRIXOILOOFDPSDLJQSUomises out of a desire to be re±elected. This would require mayors to be good politicians but I had not formally considered that good politicians do not translate into solid public managers. The ability to manage staff, change bureaucracies, work well with donors, and put in place the processes to deliver on campaign promises were conditions in Maputo, Dondo and Beira. Leadership was the link that united the use of increased revenue with donor aid and government transfers to manage improvement in municipal services. While re-election was a legitimate and prominent driver, municipal service delivery could not have occurred in the absence of mayoral leadership, a component that was missing in Matola.

How leadership affects government performance is a well-studied linkage in the democracy literature. Robert Putnam studied Italian regional government from 1970 to 1989 and observed that regional governments become more professionalized and better able to deliver services as the role of the regional councilor became more professionalized and institutionalized. One study of 390 municipalities in Latin America suggests that local executives both play a pivotal role in negotiating between central government and citizens for resources and determining the extent to which citizens have input into the provision of municipal services (Andersson and Laerhoven 2007). Another study of three Brazilian municipalities similarly points to the importance of mayoral capacity in policy implantation (Wampler

2004). One study of 40 Columbian municipalities found that mayoral educational background and job- related experience positively influenced municipal education performance (Avellaneda 2008).

,REVHUYHGWKDWOHDGHUVKLSLQWKH$IULFDQFRQWH[WWHQGVWREHYLHZHGWKURXJKWKH³ELJPDQ´OHQVLQ which leaders are generally observed to be patrimonial and self-interested, protecting their access to state resources. Bratton and Mattes did a comprehensive review of citizen attitudes in six African countries that suggested that ³ELJPDQ´DQGVLQJOH-party rule continues to be tolerated (Bratton and

Mattes 2004)$V&ODXGH$NHKDVREVHUYHG³GHPRFUDF\KDVEHHQUHGXFHGWRWKHFUXGHVLPSOLFLW\RI

PXOWLSDUW\HOHFWLRQVWRWKHEHQHILWRIVRPHRIWKHZRUOG¶VPRVWQRWRULRXVDXWRFUDWV´ $NH

However, I argued in Chapter 6 that mayoral leadership was a key factor in municipal performance improvement. I found in Mozambique that the story was more nuanced.

234

My study found that the three municipalities with high performance had three mayors who had technocratic backgrounds, could put together a team of trusted and capable advisors to implement changes in municipal performance, had ways to regularly consult their constituents, and had donor funding in support of improvements in trash collection and market management. In the case of Maputo,

Dr. Comiche was a former Governor of the Central Bank of Mozambique, a professor at Eduardo

Mondlane University, functionary within the Ministry of Finance and trained Ph.D. economist. He approached the World Bank to convert PDM funding into support for three workshops of stakeholders that eventually led to the 10-year PROMAPUTO economic and urban development strategy. He appointed staff who had varied, but related, experience in their respective areas of supervision and held them accountable with daily meetings. Through KLV³RSHQPDQGDWH´SROLF\KHPHWZLWKFRQVWLWXHQWV on a rotational basis every 2-3 months and worked through his town councilors to follow-up and report back to those neighborhoods n subsequent hearings. Under his mandate, the first participatory budget for the city of Maputo was submitted to the Municipal Assembly for FY 2008. Dr. Comiche had both the technical background and management experience to run the CMM.

In Dondo, Manual Cambezo had two, 5- year terms as mayor of the municipality and town councilors he appointed in chief positions, such as sanitation and finance, had been with him since his

ILUVWWHUP3ULRUWREHFRPLQJPD\RURI'RQGRKHKDGEHHQKHDGRI%HLUD¶V([HFXWLYH&RXQFLOWKH predecessor to the decentralized CMB. The Austrian Development CorSRUDWLRQ¶VVXSSRUWRI participatory budgeting created a feedback mechanism of regular citizen hearings that informed

'RQGR¶VDOORFDWLRQRILQIUDVWUXFWXUHIXQGV+HZDVRQHRIVL[PD\RUVQDWLRQZLGHRIDOO municipalities who served continuously since 1998, a fact recommending that he could both meet party and citizen expectations.

In Beira, Daviz Simango was a trained civil engineer and had worked on waterfront projects for the municipality as a private contractor prior to being elected Mayor As first-term RENAMO mayor in

Beira, he appointed many new staff to municipal positions from market chief to town councilor level.

Interviews with municipal managers indicated that in the area of trash collection, the Mayor and his

235

front office became personally involved, calling staff directly for updates on equipment, trash collection and routes. Mayor Simango placed a trusted advisor, his cousin, in the role of town councilor of markets and charged him with increasing fee collections, which he did. While Mayor Simango did not have routine hearings as did Dr. Comiche or a participatory budget process in place as was the case in

Dondo, I observed neighborhood leaders of some of the poorest informal neighborhoods regularly come to City Hall to consult with various ton councilors. While Mayor Simango indicated that initial improvements in trash collection occurred without donor assistance, the Italian Cooperation was active in Beira and worked closely with the CMB in providing project funds to improve trash collection. All three had technical backgrounds, familiarity with appointing and managing staff and projects and had a technical expertise that contributed to the running of their respective municipalities. In all three cases, donors had good relationships with the mayors, which facilitated the ability those mayors had to articulate cases for funding projects. These conditions did not exist in Matola due to the interrupted tenure of the mayor and appointment of interim leadership.

0D[:HEHU¶VVWXG\RIRUJDQL]DWLRQVVXJJHVWHGOHDGHUVKLSDXWKRULW\UHVWHGRQRQHRIWKUHH grounds: rational, legal rules and norms; tradition or charisma (Weber 1947). Even Putnam observed the role of charismatic leadership in Italy by documenting that more charismatic regional councilors in Italy

HOHFWHGLQWKHZKHUHODWHUUHSODFHGE\WKH³PRUHSURVDLFPDQDJHULDOW\SHV´ 3XWQDP, 24). I found that a degree of all three were present in the three cases with high performance but would recommend for future study how these interacted to better understand the organizational and cultural factors that lead to improve in governance and quality of life of residents of the municipal space. .

:KLOH'U&RPLFKH¶VOHDGHUVKLSIRFXVHGRQHVWDEOLVKLQJUXOHVDQGQRUPVZLWKLQWKH&00WKH traditional dominance of FRELIMO in both Dondo and Maputo were present. In Beira, Daviz

6LPDQJR¶VOHDGHUVKLSVW\OHZDVFKDULVPDWLFDQGWUDGLWLRQDODVHYLGHQFHGE\WKH1GDXVRQJVVXQg at the

MDM Congress and his Ndau nickname, Chinamiru CheduWKH³RQHZKRXQLWHVXV´,QERWK%HLUDDQG

Dondo, traditional forms of authority mixed with the new rules and norms of decentralization in resolving conflict. In a market in Dondo, vendors relied both on the CMD and the local chapter of the

236

WUDGLWLRQDOKHDOHU¶VDVVRFLDWLRQVRUHVROYHDGLVSXWHZLWKDYHQGRULQWKHPDUNHWZKRZDVVFDULQJRII customers. In Beira, the neighborhood secretary, appointed by the CMB, worked closely with traditional healers to resolve robberies and the drowning of a local boy. Even in Maputo, a major urban infrastructure study commissioned by the CMM concluded with the following observation of Maputo:

The traditional/modern relationship should always be taken into account when one thinks of the city. For historic factors, values and practices said to be traditional coexist with the ones called modern in the urban space, shaping the peculiar form of culture and urban life. The traditional ZHGGLQJFRH[LVWVZLWKWKHFLYLOZHGGLQJDWWKH3DODFHRIWKH:HGGLQJV7KH³WUDGLWLRQDOPHGLFLQH´ KLJKO\FRQVXPHGLQWKHXUEDQVSDFH FRH[LVWVZLWKZKDWLVFDOOHGPRGHUQPHGLFLQH«Whe formal PDUNHWFRH[LVWVZLWKWKH³LQIRUPDO´«WKH³FHPHQWFLW\´DQGWKH³FDQHFLW\´WKHIODWDQGWKHVWUDZ house. These and other aspects make this city an type of hybrid urban phenomenon, a unique PRVDLFLQWKHUHJLRQ´ Plano de Estrutura do Município de Maputo 2008, 112).

Given this linkage, an area of further study should involve examining cases of traditional or charismatic leadership in which government performance is high and how these types of leadership co-exist with

OHDGHUVKLSEDVHGRQ³UDWLRQDO´JURXQGVWRDFKLHYHWKHVHRXWFRPHV

Conclusion

In his 2010 publication Emerging Africa: How 17 Countries are Leading the Way, Steven

Radelet identified five fundamental changes in 17 African countries, Mozambique among them, have experienced in the past 15 years that have sustained economic growth, deepened democracy and reduced levels of poverty. They are: 1) more democratic accountable government; 2) more sensible economic policies; 3) end of debt crisis and change in donor relationships; 4) new technologies; and, 5) a new generation of leadership. Some of these trends are reflected at a micro-level in the findings I

GLVFXVVHGLQWKLVGLVVHUWDWLRQ0R]DPELTXH¶VGHFHQWUDOL]DWLRQUHIRUPVLQVWLWXWHGFKDQJHVLQ

SROLWLFDODQGUHYHQXHDXWKRULWLHVWKH³SHUVLVWHQWDQGFRQQHFWHGVHWRIUXOHV´ZKLFKLQFHQWLYL]HG accountability at the local level through municipal elections and performance in three of my four cases

237

(Keohane 1988, 384). The 2008 elections demonstrated how mayors in my cases built campaigns based on performance and, even in the presence or the dominance of both the FRELIMO party and

FRELIMO-RENAMO dynamic²%HLUD¶VPD\RUUDQZRQDQGFUHDWHGDQRWKHUSROLWLFDOSDUW\(YHQLQ

Maputo, where the FRELIMO mayor was not re-nominated in 2008 to run, illustrating FRELIMO dominance, voters knew that this mayor had tried to reform City Hall and news sources using the

Internet could publish articles about the nominations process. Furthermore, the mayor ran in 2003 with the incentive to fulfill his campaign promises, an incentive which did not exist pre-1998.

Civil society existed in many different forms in the four cases I studied ranging from an active lobby pressuring for concessions on municipal policies and prospective income such as ASSOSTI in

Maputo to the absence of neighborhood trash collection organizations and weak Merchant Associations in Matola. Civil society, in the form of neighborhood organizations, was fostered by donors in efforts to buttress municipal service provision post-decentralization. In some cases, such as Maputo, this led to microenterprises with private contracts from the city and in others, such as Beira, it meant an expansion and contraction of organizations dependent on donor funding. In Dondo, it meant, not the creation of organizations but rather the creation of a participatory process in which citizen feedback could be regularly funneled to municipal officials. Colonialism left social infrastructure in the form of market chiefs and socialism incubated the Merchant Associations. These Associations extended the municipalLW\¶Vreach and helped facilitate communication between municipalities and vendors, which, in some cases, resulted in better municipal service, as discussed.

Decentralization provided incentives and opportunities to municipalities to raise local revenue, even in the one case where performance was low. Across all cases, own-source municipal revenue grew from 2004 to 2007. In Maputo, relationships with donors led to sustainable service improvements and technical assistance, which directly contributed to this increase. However, the explosive growth in own- source revenue across all cases is strong evidence that suggests the incentive to use these authorities was strong enough for municipalities to pursue collection. Simultaneously, central government transfers grew and shifted towards infrastructure investment. These facts combined point to an increased investment in municipal government froPWR/LNH&ROOLHU¶VZRUNRQVFUXWLQL]HGUHYHQXH 238

further study of how unearned income positively impacts performance and why would lead us to a

EHWWHUXQGHUVWDQGLQJRIKRZVWDWHVZLWKFRQVWUDLQWVVXFKDV0R]DPELTXH¶VFDQVXFFHVVIXOO\LQVWLWXWH changes in institutions and improvements in quality of life for their citizens.

While decentralization provided opportunities to improve accountability and government

SHUIRUPDQFHDVP\UHVHDUFKVKRZHG³GHFHQWUDOL]DWLRQLVDQLQVWLWXWLRQDOPHDQVRQO\´ /LQGHU,

26). Taking a broader, national view, government performance in Mozambique faces substantial challenges. Per capita remains low and half of the population remains in poverty. However, as Radelet highlights, factors such as changing donor relationships, more sensible economic policies and more accountable government are improving living standards, reducing poverty and increasing economic growth for 300 million Africans, half the population of the continent, Mozambique included (Radelet

2010). This is not an optimistic view but one backed by both the research in this dissertation and recent scholarly and empirical work. The year 2008 marked the ten-year anniversary of decentralization in

Mozambique. A study led by the World Bank with a broad base of participants concluded:

The municipal governments of Mozambique are poised to reach the end of their first decade of existence mired by a series of complex economic constraints, growing social problems, uncertain policies and institutional weaknesses. Despite these negative factors, the last ten years saw these institutions create basic organizational structures that work. (Municipal governments) assumed the responsibility to provide basic services to their populations, manage their territories, and, more recently, leading the economic development of their cities. Now, the challenge is to attain a sustainable and integrated development process for expanding, urban areas and, in a broader sense, for the entire country (Illal and Vásconez 2009, 267).

As this dissertation has demonstrated, changes in institutions can create changes in political and revenue pressure to improve municipal serviFHGHOLYHU\$V.HRKDQHREVHUYHG³LQVWLWXWLRQVGRQRWPHUHO\ reflect the preferences and power of the units constituting them; the institutions themselves shape those

SUHIHUHQFHVDQGWKDWSRZHU´ .HRKDQH, 382).

239

Table 7.1. Hypothesis and Findings

Hypothesis: Decentralization creates incentives that motivate municipal government officials to act and opportunities for residents to pressure them.

RESOURCE PRESSURE

Revenue:

POLITICAL PRESSURE

Civil Society:

Expected Finding: Strong civil society can pressure municipal government to improve municipal services.

Elections:

Expected Finding: Competitive elections incentivize mayors to improve municipal services to constituents.

240

Table 7.2. Expected v. Actual Findings

CIVIL SOCIETY REVENUE ELECTIONS MUNICIPAL PERFOMANCE Expected Actual Expected Actual Expected Actual

MAPUTO

Markets and Strong Strong Earned Earned Medium High Trash Income Income Medium Collection

BEIRA

Markets and Strong Active Earned Earned High High High Trash Income Income Collection

DONDO

Markets and Strong Active Earned Unearned Low Low High Trash Income Income Collection

MATOLA

Markets and Weak Weak Unearned Unearned N/A Low Trash Income Income Collection

241

BIBLIOGRAPHY

"40 Por Cento Dos Habitantes Da Beira São Seropositivos", Noticias Lusófonas http://www.centrodeemergencia.com/mocambique/2005/03/40_por_cento_do.html (2009).

"A Faltering Phoenix: Mozambique's Recovery." The Economist July 10, 2010.

A. Agrawal, J. Ribot. "Accountability in Decentralization: A Framework with South Asian and West African Cases." The Journal of Developing Areas 33 (1999).

Abrahamsson, Hans; Nilsson, Anders. Mozambique, the Troubled Transition: From Socialist Construction to Free Market Capitalism. Translated by Mary Dally. London and New Jersey: Zed Books, 1995.

Adriano Nuvunga, Marcelo Mosse, and César Varela. Relatório Do Estudo Sobre Transparência, Áreas De Riscos E Oportunidades De Corrupção Em Seis Autarquias Moçambicanas: Chibuto, Chimoio, Dondo, Gurue, Nacala and Vilanculos. Maputo, 2007.

Ake, Claude. Democracy and Development in Africa. Washington, D.C.: Brookings Institute Press, 1996.

Allison, Graham T. Essence of Decision: Explaining the Cuban Missile Crisis. 1st ed. Boston: Little Brown, 1971.

Álvaro, Bernardo. "Enquanto Frelimo Vence Em 42 Das 43 Autarquias." Canal de Moçambique 2008.

Andersson, Krister; van Laerhoven, Frank. "From Local Strongman to Facilitator: Institutional Incentives for Partifcipatory Municipal Governance in Latin America." Comparative Political Studies 40, no. 9 (2007): 1085-1111.

Andrews, Matthew. Voice Mechanisms and Local Government Fiscal Outcomes: How Do Civic Pressure and Participation Influence Public Accountability? Public Expenditure Analysis: Public Sector Governance and Accountability Series, Edited by Anwar Shah. Washington, D.C.: The International Bank for Reconstruction, 2005.

Anexo 2006-2007. Ministry of Planning and Finance. Maputo, Mozambique. 2007.

Anexo 2008. Ministry of Finance. Maputo, Mozambique. 2008.

Arias, Enrique Desmond. "Faith in Our Neighbors: Networks and Social Order in Three Brazilian Favelas." Latin American Politics and Society 46, no. 1 (Spring 2004): 1-38.

"ASSOTI President." Maputo, Mozambique: Interview by author, 2008.

242

Avellaneda, Claudia N. "Municipal Performance: Does Mayoral Quality Mater?" Journal of Public Administration Research and Theory (2008): 285-313.

Aworti, Nicolas. "The Past, Present, and Future of Decentralisation in Africa: A Comparative Case Study of Local Government Development Trajectories of Ghana and Uganda." International Journal of Public Administration 33, no. 12/13 (2010): 630-631.

Azagaia, Mano. "Para Quem Duvidava Do Poder Do Povo." O País, Maputo, Mozambique. 2008, 14.

Bahl, Roy. "Fiscal Decentralization and Development." Public Budgeting & Finance (Summer 1999): 60-61.

Bahl, Roy and Johannes Linn. "Fiscal Decentralization and Intergovernmental Transfers in Less Developed Countries." Publius 24, no. 1 (Winter 1994): 1-19.

Baiocchi, Gianpaolo. "Participation, Activism and Politics: The Porto Alegre Experiment and Deliberative Democracy Theory." Politics and Society 29 (2001): 43-71.

Bardhan, Pranab. "Decentralization of Governance and Development." Journal of Economic Perspectives 16, no. 4 (Fall 2002): 202.

Bardhan, Pranab and Dilip Mookherjee. "Capture and Governance at Local and National Levels." American Economic Association 90, no. 2 (2000): 135-139.

-- "Decentralisation and Accountability in Infrastructure Delivery in Developing Countries." The Economic Journal 116 (January 2006): 101-127.

Barnes, Dr. John. "Descentralização Em Moçambique: Mito Our Realidade?". Maputo: CTA, Program of Planning and Decentralized Finance, Directorate of National Planning and the Budget.

Bates, Robert; Lien, Da-Hsiag D. "A Note on Taxation, Development, and Representative Government." Politics and Society 14 (1985): 53-70.

Batley, Richard, Liv Bjørnestad, and Amélia Cumbi, "Mozambique Country Report". Joint Evaluation of General Budget Support: 1994±2004 Burkina Faso, Malawi, Mozambique, Nicaragua, Rwanda, Uganda, Vietnam. International Development Department School of Public Policy University of Birmingham, U.K.. May 2006. Accessed on September 22, 2011 at http://www.usaid.gov/mz/doc/misc/moz_eval_budget%20support.pdf

"Beira Já Tem Meios Para Recolha De Lixo." In Moçambique para todos, 2006.

Bernard, H. Russell. Research Methods in Anthropology: Quantitative and Qualitative Approaches. 4th ed. Lanham, MD: Altamira Press, 2006.

Bird, Richard. Intergovernmental Fiscal Relations in Latin America: Policy Design and Policy Outcomes. Washington, D.C.: Inter-American Development Bank, 2000.

Bird, Richard; Vaillancourt, François. "Fiscal Decentralization in Developing Countries: An Overview." In Fiscal Decentralization in Developing Countries, edited by Richard; Vaillancourt Bird, François. Cambridge, MA: Cambridge University Press, 1998.

243

Bonnal, Jean, "History of Decentralization" http://www.ciesin.org/decentralization/English/General/history_fao.html (accessed December 17 2010).

Brancati, Dawn. "Decentralization: Fueling the Fire or Dampening the Flames of Ethnic Conflict and Secessionism?" International Organization 60, no. 3 (2006): 651-685.

%UDWWRQ0LFKDHO0DWWHV5REHUW:KDW³WKH3HRSOH´6D\$ERXW5HIRUPV,QDemocratic Reform in Africa: The Quality of Progress, edited by E. Gyiman-Boadi. London and Boulder: Lynne Rienner Publishers, 2004.

Bräutigam, Deborah. "Introduction: Taxation and State-Building in Developing Countries." In Taxation and State-Building in Developing Countries: Capacity and Consent, edited by Deborah; Fjeldstad Bräutigam, Odd-Helge; Moore, Mick, 1-33. Cambridge: Cambridge University Press, 2008.

-- "The People's Budget? Politics, Participation and Pro-Poor Policy." Development Policy Revi ew 22, no. 6 (2004).

--. Aid Dependence and Governance. Stockholm: Expert Group on Development Issues, 2000.

-- "Institutions, Economic and Democratic Consolidation in Mauritius." Comparative Politics 30, no. 1 (Oct. 1997): 45-62.

Bräutigam, Deborah; Knack, Stephen. "Foreign Aid, Institutions, and Governance in Sub-Saharan Africa." Economic Development and Cultural Change 52, no. 2 (January 2004): 266.

Cabrita, João. The Tortuous Road to Democracy. New York: St. Martin's Press, 2000.

Carbone, Giovanni M. "Continuidade Na Renovação? Ten Years of Multiparty Politics in Mozambique: Roots, Evolution and Stabilization of the Frelimo0-Renamo Party System." Journal of Modern African Studies 43, no. 3 (2005): 422.

Cem Anos Da Beira. 2007.

Census 2007. National Institute of Statistics. Maputo, Mozambique. 2007.

Chabal, Patrick; Daloz, Jean-Pascal. Africa Works: Disorder as Political Instrument. London and Bloomington: International African Institute in association with James Currey, Oxford, 1990.

Chamusse, Anastácia Sebastião. "A Assistência Do Banco Mundial Ao Desenvolvimento De Moçambique. Uma Análise Dos Projectos De Investimento." Thesis, ISCTE, 2008.

&KDPXVVH-RDR HG 5HXQLmR'D8QLmR'HPRFUDWD,QWHUQDFLRQDO(P3DULV³21RVVR3UREOHPD(P Moçambique É De Liderança"." Canal de Moçambique (June 30, 2008).

Cheema, G. Shabbir. "Devolution with Accountability: Learning from Good Practices." In Decentralizing Governance: Emerging Concepts and Practices, edited by G Shabbir Cheema and Dennis A. Rondinelli, 171. Washington, D.C.: Brookings Institute Press, 2007.

-- Press Release: U.S. Department of State Invites U.S. Private Sector Engagement in World Urban Forum 2010, 2010.

244

Cheema, G. Shabbir and Rondinelli, Dennis A. "From Government Decentralization to Decentralized Governance." In Decentralizing Governance: Emerging Concepts and Practices, edited by G. Shabbir Cheema and Dennis A. Rondinelli. Washington, D.C. and Cambridge, MA: Brookings Institution Press and Harvard University John F. Kennedy School of Government Ash Institute for Democratic Governance and Innovation, 2007.

Chichava, Sérgio. "MDM: A New Political Force in Mozambique?" In Election processes, liberation movements and democratic change in Africa0DSXWR,QVWLWXWRGH(VWXGRVVRFLDLVH (FRQyPLFRV ,(6(

--. Movimento Democrático De Moçambique: Uma Nova Força Política Na Democracia Moçambicana? Maputo: Instituto de Estudos Sociais e Económico (IESE) Cadernos 2010.

Cholera Drama in Beira. AIM, Maputo, Mozambique. 1998.

Cistac, Gilles. Manual De Direito Das Autarquias Locais0DSXWR/LYUDULD8QLYHUVLWiULD8QLYHUVLGDGH Eduardo Mondlane, 2001.

Collier, David, editor. The New Authoritarianism in Latin America. Princeton: Princeton University Press, 1979.

City of Biera (CMB) Contas De Gerência. 2004-2007.

City of Dondo (CMD) Contas De Gerência. 2004-2007.

City of Maputo (CMM) Contas De Gerência. 2004-2007.

Collier, David; Mahoney, James; Seawright, Jason. "Claiming Too Much: Warning About Selection Bias." In Rethinking Social Inquiry: Diverse Tools, Shared Standards, edited by Henry E; Collier Brady, David, 85-102. Lantham, MD: Rowman & Littlefield Publishers, Inc., 2004.

Collier, Paul. "Is Aid Oil? An Analysis of Whether Africa Can Absorb More Aid." World Development 34, no. 9 (2006): 1482-1497.

Collier, Paul; Dehn, Jan. "Aid, Shocks, and Growth." World Bank Policy Research Paper (2001): 1-21.

Comiche, Dr. Eneas. Interview by author, Maputo, Mozambique. December 12, 2009.

Comiche Promises Municipal Ombudsmen. AIM, Maputo, Mozambique, 2003

Crook, Richard C.; Manor, James. Democracy and Decentralisation in South Asia and West Africa: Participation, Accountability, and Performance. Cambridge: Cambridge University Press, 1998.

Crook, Richard C.; Sverrisson, Alan S. "Does Decentralization Contribute to Poverty Reduction." In Changing Paths: International Development and the New Politics of Inclusion, edited by Peter P. Houtzager and Mick Moore, 252. Ann Arbor: University of Michigan Press, 2008.

Cuereneia, Aiuba. The Process of Decentralisation and Local Governance in Mozambique: Experiences and Lessons Learnt. Cape Town: United Nations Capital Development Fund, 2001.

Cox, Robert. Approaches to World Order. Cambridge: Cambridge University Press, 1996.

245

Dahl, Robert. Polyarchy: Participation and Opposition. New Haven: Yale University Press, 1971.

"Daviz Simango Exonera Três Vereadoes", Savana http://www.imensis.co.mz/ (2010).

De Brito, Miguel. EISA, Interview with author, Maputo, Mozambique 2009.

"Decentralização Em Moçambique: Mito Ou Realidade?". Maputo: Program of Planning and Decentralized Finance. de Renzio, Paolo. "Budget Reforms in Mozambique: Towards a Political Economy Approach,." In IESE Conference on Challenges for Social and Economic Research in Mozambique. Maputo, Mozambique, 2007.

-- "Aid, Budgets and Accountability: A Survey Article." Development Policy Review 24, no. 6 (2006): 627-645.

DeSoto, Hernando. The Mystery of Capital: Why Capitalism Triumphs in the West and Fails Everywhere Else. New York: Basic Books, 2000.

Devarajan, Shantayanan; Rajkumar, Andrew S.; Swaroop, Vinaya. What Does Aid to Africa Finance? Washington, D.C.: The World Bank, 1999.

Diamond, Larry. "Civil Society and the Development of Democracy." Estudio/Working Paper 1997/101 June 1997, accessed on September 24, 2001 at http://www.march.es/ceacs/ingles/publicaciones/working/archivos/1997_101.pdf

--Developing Democracy: Toward Consolidation. Baltimore: Johns Hopkins University Press, 1999.

Diamond, Larry, and Juan Linz, and Seymour Lipset. Democracy in Developing Countries Africa. Boulder: Lynne Reiner Publishers, 1988.

Did Frelimo Candidate Steal State Property? : AIM. Maputo, Mozambique. 2003.

"Directory of Development Organizations: Guide to International Organizations, Governments, Private Sector Development Agencies, Civil Society, Universities, Grantmakers, Banks, Microfinance Institutions and Development Consulting Firms", Resource Guide to Development Organizations and the Internet http:/www.devdir.org.

Dondo: Dois Candidators Rezam AntHV'R(QFDUUDPHQWR'Dµ&DoD¶. Society and Politics, 2008.

Eaton, Kent, Tyler Dickovick, and Paul Smoke, USAID Democratic Decentralization Programming Handbook, ARD, Inc. Contract AEP-I-00-00-00016-00. June, 2009, accessed online September 24, 2011 at http://www.usaid.gov/our_work/democracy_and_governance/publications/pdfs/DDPH_09_22_ 09_508c.pdf.

Eckardt, Sebatian. "Political Accountability, Fiscal Conditions and Local Government Performance- Cross-Sectional Evidence Fom Indonesia." Public Administration and Development 28, no. 1 (2008).

"Electricidade De Moçambique", http://www.edm.co.mz/ (accessed May 16 2011).

246

"Eneas Comiche Negou Terrenos a Camaradas." In Moçambique para todos blog: Savana, 2008.

Eubank, Nicolas. "Taxation, Political Accountability, and Foreign Assistance: Lessons from Somaliland." In SSRN Wor ki ng Paper , 2010.

Faguet, Jean-Paul. Governance from Below in Bolivia: A Theory of Local Government with Two Empirical Tests. Berkeley: Center for Latin American Studies at University of California, Berkeley, 2009.

Faria, Fernanda; Chicava, Ana. Descentralização E Cooperação Descentralizada Em Moçambique. International Labor Organization, October 1999.

Fatton, Robert. "Africa in the Age of Democratization: Civic Limitations of Civil Society." African Studies Review 38, no. 2 (September 1995): 69.

Fauvet, Paul. Maputo: Information Agency of Mozambique (AIM), Interview by author, 2009.

Fay, Marianne and Chaarlotte. Urbanization without Growth: A Not-So-Uncommon Phenomenon. World Bank, 2000.

FDC Mozambican Civil Society Within: Evaluation ,Challenges, Opportunities and Action. United Nations Development Programme, 2007.

Fengler, Wolfgang, "Can Rapid Population Growth Be Good for Economic Development?", World Bank http://blogs.worldbank.org/africancan/can-rapid-population-growth-be-good-for- economic-development.

Figuerieda, Anselmo. Dondo: Interview by author, Town Councillor for Finances and Municipal Property, November 7, 2008.

Fjeldstad, Odd-Helge. "Taxation, Coercion and Donors: Local Government Tax Enforcement in Tanzania." Journal of Modern African Studies 39, no. 2 (2001): 301.

--"What's Trust Got to Do with It? Non-Payment of Service Charges in Local Authorities in South Africa." Journal of Modern African Studies 42 (2004): 539-562.

Folha Informativa Dos Municípios II. Ministry of State Adminsitration, Maputo, Mozambique, 2002.

)RUTXLOKD6DOYDGRU&DGHWH³5HPHQGR1RYR(P3DQR9HOKR´2,PSDFWR'DV5HIRUPDV'H Descentralização No Processo De Governação Local Em Moçambique." In Challenges for the Social and Economic Investigation in Mozambique, 1-2: Institute of Social and Economic Studies Inaugural Conference, 2007.

Francisco, António; Mucavele, Albertina; Monjane, Paul; Seuane, Sónia. Índice Da Sociedade Civil Em Moçambique 2007: A Sociedade Civil Moçambicana Por Dentro: Avaliação, Desafios, Oportunidades E Acção Foundation for Community Development 2008. Maputo, 2008.

)UHOLPRµ&DSLWXOD¶¬9LWyULD'H'DYL]6LPDQJR1D%HLUD,Q Macau Blog, 2010: Canal de Moçambique, 2000.

Frelimo Chooses Its Sofala Candidates. AIM. Maputo, Mozambique. 2008. 247

Frelimo Promises Battle against Corruption. AIM. Maputo, Mozambique, 2003.

Frelimo Striving for Strong Municipal Candidates. AIM. Maputo, Mozambique. 2008.

Fukuyama, Francis. The End of History and the Last Man. New York: Avon Book, 1993.

Fung, Archon; Wright, Erik. "Deepening Democacy: Innovations in Empowered Participatory Governance." Politics and Society 29 (2001): 5-42.

Gramsci, Antonio. Selections from the Prison Notebooks of Antonio Gramsci: International Publishers, 1971.

Grindle, Merilee S. . "Local Governments That Perform Well." In Decentralizing Governance: Emerging Concepts and Practices, edited by G. Shabbir; Rondinelli Cheema, Dennis A., 56- 74. Washington, D.C.: Brookings Institute, 2007.

Guess, George. "Comparative Decentralization Lessons from Pakistan, Indonesia, and the Philippines." Public Administration Review 65, no. 2 (2005): 217-230.

Guevane, Luís. "O Efeito-Beira Como Motivador Da Ida Ás Urnas." Savana 2008, 12.

Gyimah-Boadi, E. . "Civil Society in Africa." Journal of Democracy 33, no. 12/13 (1996): 118-132.

Hanlon, Joseph. Mozambique, Revolution under Fire. London: Zed Books, Ltd, 1990.

-- Who Calls the Shots? London: James Currey, Ltd, 1991.

-- Mozambique Peace Process Bulletin 21. 1998.

--"Landslide Victory for Frelimo and Armando Guebuza." Mozambique Peace Process Bulletin, no. 43 (2009).

--. "Mozambique Political Process Bulletin." no. 43 (2009).

-- "CNE Admits Error in Key Secret Document." Mozambique Peace Process Bulletin, no. 44 (2010): 4.

Hatry, Harry. Performance Measurement: Getting Results. Washington, D.C.: Urban Institute Press, 2007.

Heller, Patrick. "Moving the State: The Politics of Democratic Decentralization in Kerala, South Africa and Porto Allegre." Politics and Society 29, no. 131 (2001): 131-163.

Hoffman, Barak. "Political Accountability at the Local Level in Tanzania." Dissertation, University of California, San Diego, 2006.

-- "The Market for Local Government in South Africa ". New York: African Studies Association, 2007.

Huntington, Samuel. The Third Wave. Norman: University of Oklahoma Press, 1991.

248

Illal, Abdul. "National Seminar on Decentralization." Maputo: Ministry of State Administration and the World Bank, 2008.

Illal, Abdul; Vásconez, Jaime "Chapter 9. Lições Aprendidas." In 'HVHQYROYLPHQWR0XQLFLSDO(P 0RoDPELTXH$V/Lo}eV'D3ULPHLUD'pcada, 267. Washington, D.C.: World Bank, 2009.

µ,¶PD'LFWDWRU¶%RDVWV'KODNDPDAIM , Maputo, Mozambique. 2008.

Implementation Completion and Results Report (Ida-35490 Ida 3549a) on a Credit in the Amount of Us$33.6 Million to the Republic of Mozambique for the Municipal Development Project. World Bank, 2007.

Independent Campaigns Almost Invisible. AIM, Maptuo, Mozambique. 2003.

Informe Trimestral Sobre a Epidemia De Cólera Em Moçambique 2010.

Isaacman, Allen and Barbara. Mozambique: From Colonialism to Revolution, 1900-1982. Boulder: Westview Press, 1983.

Johnson, Craig; Deshingkar, Priya; Start, Daniel. "Grounding the State: Devolution and Development in ,QGLD¶V3DQFKD\DWVJournal of Development Studies 41, no. 6 (2005): 937-970.

Key Competencies for Improving Local Governance: Concepts and Strategies. UN Centre for Human Settlements, 2005.

Keohane, Robert. "International Institutions: Two Approaches." International Studies Quarterly 42, no. 1 (1998): 383.

Kim, Soonhee. "Public Trust in Government in Japan and South Korea: Does the Rise of Critical Citizens Matter?" Public Administration Review (September/October 2010): 801-810.

King, Gary; Keohane, Robert O.; Verba, Sidney. Designing Social Inquiry: Scientific Inference in Qualitative Research: Princeton University Press, 1994.

Kjær, Anne Mette. ³Institutional history or quid-pro-quo? Exploring revenue collection in two Ugandan districts paper prepared for delivery at, the 2004 Annual APSA Meeting, September 2 -5, 2004, Chicago. Aarhus: Department of Political Science, University of Aarhus.

-- "Decentralization, Accountability, and Local Government Tax Collection: A Case Study from Uganda." In The Frontiers of Political Decentralization, IDS, RUC, 2005.

Komai, Janos. The Socialist System: The Political Economy of Communism. Princeton: Princeton University Press, 1992.

Kruks-:LVQHU*DEULHOOH$IWHUWKH)ORRG&ULVLV9RLFHDQG,QQRYDWLRQLQ0DSXWR¶V6ROLG:DVWH Management." MIT, 2006.

Kulipossa, Fedilx. "Decentralisation and Democracy in Developing Countries: An Overview." Development in Practice 14, no. 6 (2004).

Lambright, Gina. Decentralization in Uganda: Explaining Successes and Failures in Local Governance. Boulder: Lynne Rienner Publishers, 2011.

249

Lewis, Peter M. "Political Transition and the Dilemma of Civil Society in Africa." Journal of International Affairs 46, no. 1 (1992): 31-54.

-- Growing Apart: Oil, Politics, and Economic Change in Indonesia and Nigeria. Ann Arbor: University of Michigan Press, 2007.

Lijphart, Arend. Patterns of Democracy: Government Froms and Performance in Thirty-Six Countries. New Haven and London: Yale University Press, 1999.

Linder, Wolf. "On the Merits of Decentralization in Young Democracies." Publius: The Journal of Federalism 40, no. 1 (2009): 25.

Lista das Micro-Empresas, Recolha Primária nas Zonas Suburbanas, CMM, Direcção de Planificação e Monitoramento, Maputo, Mozambique.

Lundin, Iraê Baptista "Citizen Political Participation in a Changing Society: The Challenge of Abstentions in the 2008 and 2009 Elections in Mozambique " (2010).

Makumbe, John. "Is There a Civil Society in Africa?" International Affairs 74, no. 2 (1998): 305-317.

Manguana, Celso. ³'RFKXPERGH&RPLFKHDRFRQFHLWR GHEDVHV³Canal de Moçambique, August 27, 2008, accessed at http://macua.blogs.com/moambique_para_todos/2008/08/do-chumbo-de-co.html on August 15, 2011.

Manndorff, Hannes; Nöst, Barbara; Reiter, Walter; dos Santos José, Ester Evaluation of the Country Programme Mozambique of the Austrian Development Cooperation: Final Report. Vienna: Lechner, Reiter und Riesenfelder Sozialforschung OEG, 2004.

Manning, Carrie. "Assessing African Party Systems after the Third Wave." Party Politics 11, no. 6 (2005).

0DQXDO%LVVRSR1Dµ&DoD¶$R9RWR. Society and Politics, 2008.

Mapa De Controlo Do Rendimento Diario De Mercados. CMD, 2006.

Marques, Lino. "Ilusão De Sucesso." Savana, December 14, 2007, 3.

Marrengula, Ramos. Maputo: Interview by author, 2008.

Mate, CaPLOR6XEVLVWHP&RPLFK}HV«Zambeze 2008, 14.

Mattes, Robert; Shenga, Carlos. µ8QFULWLFDO&LWL]HQVKLS¶LQDµ/RZ,QIRUPDWLRQ¶6RFLHW\0R]DPELFDQV in Comparative Perspective. December 2007.

Mavila, António. "Pre-Campanhã Da Frelimo Atrapalha Oposição." (2008): 25.

Mazula, Brazão. "Mozambique: The Challenge of Democratization." In Democratic Reform in Africa, edited by E. Gyiman-Boadi, 183-200. Boulder: Lynne Reiner Publishers, 2004.

Mayor of Beira Expelled from Renamo. AIM, 2008.

250

"Mendicidade E Crime São Preocupantes Em Maputo-Artus Cananá À Margem Das Celebrações Do Dia Da Cidade Ontem Assinalado." In Noticias, 2010: Maputo.co.mz, 2003.

"Moçambique: Feitiços, Cobras E Lagartos! Crónicas Romanceadas." In Moçambique Editora, 2005.

Moe, Terry. "Power and Political Institutions." Perspectives on Politics 3, no. 2 (June 2005): 215-233.

Moore, Mick. "Death without Taxes: Democracy, State Capacity, and Aid Dependence in the Fourth World." In The Democratic Developmental State: Politics and Institutional Design, edited by Mark; White Robinson, Gordon. Oxford: Oxford University Press, 1998.

-- "Revenues, State Formation and the Quality of Governance in Developing Countries." International Political Science Review 25, no. 3 (2004): 297-319.

-- "Between Coercion and Contract: Competing Narratives on Taxation and Governance." In Taxation and State-Building in Developing Countries: Capacity and Consent, edited by Deborah; Fjeldstad Bräutigam, Odd-Helge; Moore, Mick, 34. Cambridge: Cambridge University Press, 2008.

0RTXLYDODND'HGp4XH2&RPLFKH'LWRX'HUURFDGD+XPLOKDQWH1DVµ,QWHUQDV¶'D)UHOLPR",Q MiradourOnline, 2008.

Moreno-Jaimes, Carlos. "Do Competitive Elections Produce Better-Quality Governments?" Latin American Research Review 42, no. 2 (2007): 136-153.

Mosse, Marcelo. A Corrupção No Sector Da Justiça Em Moçambique: Documento De Discussão No. 3. Maputo: Center for Public Integrity, 2006.

Moyana, Salomão. Political Analysis of Mozambique: Final Report. Maputo: Embassy of Ireland, 2005.

Mozambique National Human Development Report 2005: Human Development to 2010-Reachi ng for the Millennium Development Goals. United Nations Developemnt Program , Maputo, Mozambique, 2006.

Mozambique Receives Us $30 Million for Promaputo Municipal Development Program. World Bank, 2007. Press Release.

"Mozambique: Improving Municipal Governance in Dondo, Case Study 20", Making Planning Work. 2010.

"Mozambique: Simango Publishes Documents on Beira Houses." In allAfrica.com, 2010

Mulando, Flavio. "O Papel Das Organizações Da Sociedade Civil Na Formulação De Políticas Públicas Em Moçambique: Caso Do G20 E O Parpa Ii." In Inaugural Conference of the Institute for Social and Economic Studies (IESE) Challenges for Social and Economic Investigation in Mozambique, September 19, 2007.

Munck, Gerardo L. "Tools for Qualitative Research." In Rethinking Social Inquiry: Diverse Tools, Shar ed St andar ds, edited by Henry E; Collier Brady, David, 105-121. Lantham, MD: Rowman & Littlefield Publishers, Inc., 2004.

251

Municipal Development in Mozambique: Lessons from the First Decade-Volume 1 Synthesis Report. 2009. pt. 47876-MZ.

Municipal Elections in Full. AIM, Maputo, Mozambique. 2003.

"Na Cidade De Maputo: Taxas Dividem Município E Informais", Notícias http://www.jornalnoticias.co.mz/pls/notimz2/berwsea0.simples.

"Não É Este O País Pelo Qual Negociei E Sonhei", O Pais http://www.opais.co.mz/index.php/politica/63-politica/2323-qnao-e-este-o-pais-pelo-qual- negociei-e-sonheiq.html (2010).

Natal, Isaías. "Apoiantes De Daviz Simango E Manuel Pereira Em Guerra." Magazine Independente (2008): 5.

Ndegwa, Stephen. Decentralization in Africa: A Stocktaking Survey. World Bank Africa Public Sector Reform and Capacity Building Unit, 2002.

Nhamirre , B. and E.Sambo, ³&DQGLGDWXUDGH'DYLG6LPDQJRQmRpDPHDoD"´&DQDOGH Moçambique, July 15, 2008, Maputo.

Nhamtumbo, Noé. "África Moçambique: Sem Coerência Não Há Boa Governação." CIEDIMA (2008): 21.

Nhantumbo, Noé. África Moçambique: Sem Coerência Não Há Boa Gover nação. Maputo: CIEDIMA, 2008.

North, Douglas; Weingast, Barry R. "Constitutions and Commitment: The Evolution of Institutions Governing Public Choice in Seventeenth-Century England." Journal of Economic History 49 (1989): 803-832.

"Not Quite as Stellar as It Looks: Mozambique." The Economist December 1, 2007.

Nuvunga, Adriano. "Candidatos Sem Apoio Dos Antecessores?" Boletim sobre o processo político em Moçambique 29 (2004).

--"Aid Dependence and Governance in Mozambique." In Foreign Aid, Governance and Institutional Development in Mozambique, edited by N. Aworti & A. Nuvunga, 39. Maastricht, 2007.

-- "Impact of Foreign Aid on Local Governance: A Study of Chibuto and Dondo Municipalities." In Foreign Aid, Governance and Institutional Development in Mozambique, edited by Nicolas; Adriano Aworti, Nuvunga, 54-71. Maastricht: Shaker Publishing, 2007.

Nuvunga, Adriano; Mosse, Marcelo; Varela, César. Relatório Do Estudo Sobre Transparência, Áreas De Riscos E Oportunidades De Corrupção Em Seis Autarquias Moçambicanas: Chibuto, Chimoio, Dondo, Gurue, Nacala and Vilanculos. Maputo: Center for Public Integrity, 2007.

O'Donnell, Guillermo A. Modernization and Bureaucratic-Authoritarianism: Studies in South American Politics. Vol. 9. Berkeley: Institute of International Studies, University of California, 1973.

"Odadata Glossary", http://www.odamoz.org.mz/ (2010).

252

Opinion Polls Show Comiche Most Popular Candidate. AIM, 2008.

"Os Hábitos De Mugabe²Uma Estratégia Ou Uma Ilegalidade." In TribunaFax, 2010, 2008.

"O Presidente Do Município Da Beira, Daviz Simango, Foi Hoje Expulso Da Renamo, Principal Partido Da Oposição Moçambicana, Por Se Candidatar Como Independente Às Eleições Autárquicas De Novembro, Anunciou O Conselho Nacional Do Partido", http://ww1.rtp.pt/noticias/?article=149633&visual=3&layout=10 (2011).

"Palmeiras I." In Focus group findings. Biera, 2008.

3DUWLGR)UHOLPR5HLWHUD([SHFWDWLYH'H9LWyULD1DV$XWDUTXLDV'H6RIDOD¶. Society and Politics, 2008.

Pereira, João. " Mecanismo Para Apoio Da Sociedade Civil." Maputo: Interview by author, 2008.

Pereira, João. "Partisan Identification in Africa: An Exploratory Study in Mozambique." University of Cape Town, November 19, 2008.

Plano Director: Gestão De Resíduos Sólidos Urbanos Na Cidade De Maputo. CMM, 2008.

PMR Online 18, no. 2 (2008).

Posse, Paulino Adolfo, "How Well Is Your Town Run? Mayors from across the Continent Are Meeting in Nairobi This Week at Africities 2006. How Successful Is Your Local Government?", BBC http://news.bbc.co.uk/2/hi/africa/5349562.stm.

"President of the National Association of the Municipalities of Mozambique." Matola: Interview by author, 2009.

Putnam, Robert. Making Democracy Work: Civic Traditions in Modern Italy. Princeton: Princeton University Press, 1993.

Radelet, Steven. Emerging Africa: How 17 Countries Are Leading the Way. Washington, D.C.: Center for Global Development, 2010.

Reaud, Beatrice and Weimer, Bernhard. Mozambique Decentralization Assessment. Report prepared for the United States Agency for International Development under ARD, Inc. Contract: DSD-I- 00-04-00227-00, Task Order 15. June, 2010.

Rogowski, Ronald. "How Inference in the Social (but Not the Physical) Sciences Neglects Theoretical Anomaly." In Rethinking Social Inquiry: Diverse Tools, Shared Standards, edited by Henry E; Collier Brady, David, 77-82. Lanham, MD: Rowman & Littlefield Publishers, Inc., 2004.

Roque, Carlos; Tengler, Hemma. Dondo No Dhondo: Perspectivas De Desenvolvimento Municipal Participativo. Beira: European Union, Austrian Cooperation, and Austrian Institute for North- South Cooperation, Center for Services of Sofala, 2000.

Ross, Michael. "Does Oil Hinder Democracy?" World Politics 53, no. 3 (2001): 325-361.

--"Does Taxation Lead to Representation?" British Journal of Political Science 34, no. 2 (2004): 229-249.

Rousseau, Jean Jacques. The Social Contract, 1762.

253

5XLJURN,QJH0R]DPELTXH¶V*HQHUDO(OHFWLRQVAfrican Security Review 14, no. 1 (2005).

Saide, Momade; Nhata, Virgulino. "Mozambique: Poverty Observatory and the Program Aid Partnership (Pap): Policy Dialogue, Civil Society Consultation and Mutual Accountability of Donors and Governments." In Sour cebook, 76: OECD/DAC Joint Venture on Managing for Development Results and the World Bank, 2007.

Sambo, Emildo, "Mercados Da Cidade De Maputo: Cobranças Paralelas E Coercivas Deixem Vendedores Com Nervos À Flor Da Pele", Canal de Moçambique http://www.canalmoz.com/default.jsp?file=ver_artigo&nivel=0&id=&idRec=4356 ( ).

Shenga, Carlos. Economic Conditions, Living Conditions and Poverty in Mozambique. 2010.

Simango, Daviz. "Closing Remarks at MDM Congress." Catholic University of Beira, 2008.

Simango, Daviz. Biera: Interview by author, 2009.

Sitoe, Eduardo ; Matsimbe, Zefanias; Pereira, Amilcar Parties and Political Development in Mozambique. EISA, 2005.

Stretz, Joaquim. Maputo: AGRESU, Interview by author, 2008.

Tarrow, Sidney. "Bridging the Quantitative-Qualitative Divide." In Rethinking Social Inquiry: Diverse Tools, Shared Standards, edited by Henry E. Brady; David Collier, 174. Lantham, MD: Rowan & Littlefield Publishers, Inc., 2004.

Thompson, Carol. "Beyond Civil Society: Child Soldiers as Citizens of Mozambique." African Political Economy no. 80 (1999): 195.

"Technician LQWKH6DQLWDWLRQDQG6DOXEULW\7RZQ&RXQFLORU¶V2IILFH0DWROD,QWHUYLHZE\DXWKRU 2008.

Tilly, Charles. Coercion, Capital and European States, Ad 990-1992. Cambridge, MA: Blackwell, 1992.

Timmons, Jeffrey F. "The Fiscal Contract: States, Taxes, and Public Services." World Politics 57, no. 4 (2005): 530-567.

Tocqueville, Alexis de. Democracy in America, Edited by Isaac Kramnick: Penguin Classics, 2003.

Três Candidatos Requalificados No Dondo, Gorogosa E Manica. Society and Politics, 2008.

UNICEF, "Mozambique at a Glance: A Nation in Progress" www. unicef.org/Mozambique/overview.html (2010).

Vaillancourt, François. "Fiscal Decentralizationin Developing Countries: An Overview." In Fiscal Decentralization in Developing Countries, edited by Richard; Vaillancourt Bird, François. Cambridge: Cambridge University Press, 1998.

Vásconez, Jaime (Centro Internacional de Gestion Urbana, CIGU-Ecuador) and Abdul Ilal (IPAM). ³&KDSWHU/Lo}HV$SUHQGLGDV´'HVHQYROYLPHQWR0XQLFLSDO(P0RoDPELTXH$V/Lo}HVGD 3ULPHLUD'pFDGD, 267, accessed on September 18, 2011 at http://siteresources.worldbank.org/INTMOZAMBIQUE/Resources/MunicipalPort.pdf 254

Vines, Alex. Renamo: Terrorism in Mozambique. London and Bloomington, IN: University of York with James Currey, Indiana University Press, 1991.

Wampler, Brian. "Expanding Accountability through Participatory Institutions: Mayors, Citizens, and Budgeting in Three Brazilian Municipalities." Latin American Politics and Society 46, no. 2 (Summer 2004): 73-99.

Waty, Teodoro Andrade. Autarquias Locais: Legislação Fundamental. Maputo: ML Graphics, 2000.

Weber, Max. Theory of Social and Economic Organization, Edited by Talcott Parsons. New York: Free Press, 1947.

Weimer, Bernhard; Fandrych, Sabine. "The Administrative Reform in Mozambique: A Contribution to Peace and Democracy?" Afrika Spectrum 32, no. 2 (1997).

Weingast, Barry. "Second Generation Fiscal Federalism: The Implications of Fiscal Incentives." Journal of Urban Economics, no. 65 (2009): 279-293.

Wekwete, K.H. "Decentralization Requires Local Organization." In From Government Decentralization to Decentralized Governance. Washington, D.C.: Brookings Institution, 2007.

Wekwete, K.H. "Decentralizations to Promote Effective and Efficient Pro-Poor Infrastructure and Service Delivery in the Least Developed Countries." In Decentralizing Governance: Emerging Concepts and Practices, edited by G. Shabbir; Rondinelli Cheema, Dennis A., 242. Washington, D.C.: Brookings Institute, 2007.

West, Harry G. Kupilikula: Governance and the Invisible Realm in Mozambique. Chicago: University of Chicago Press, 2005.

:LOVRQ%HQMDPLQ$QWyQLR6DORPmR$UWXU&DQDQD³)DUWR´'RV3UREOHPDV'HMaputo, Desabafa: Alguma Coisa Está Errada Entre Nós E Os Munícipes." Domingo (2001). http://www.ilo.org/public/portugue/region/eurpro/lisbon/pdf/assotsi.pdf [accessed 2010].

Woods, Dwayne. "Civil Society in Europe and Africa: Limiting State Power through a Public Sphere." African Studies Review 35, no. 2 (1992): 77-100.

World Bank. ³,PSOHPHQWDWLRQ&RPSOHWLRQDQG5HVults Report (IDA-35490 IDA 3549A) on a Credit in the Amount of US$33.60 million (SDR 26.70 million credit) to Republic of Mozambique for the 0XQLFLSDO'HYHORSPHQW3URMHFW´. Report No. ICR000072, AFT: Water and Urban 1, Southern Africa Country Cluster 2, AFRICA, August 27, 2007.

Yin, Robert. Applications of Case Study Research. 2nd ed. Thousands Oaks, CA; London, UK; New Delhi, India: Sage Publications, 2003.

255